fix to avoid loop

fix https://github.com/chrislusf/seaweedfs/issues/1966

avoid shared readErr variable

.github/workflows/release.yml

@@ -40,6 +40,7 @@ jobs:
           goarch: ${{ matrix.goarch }}
           release_tag: dev
           overwrite: true
+          pre_command: export CGO_ENABLED=0
           build_flags: -tags 5BytesOffset # optional, default is
           ldflags: -extldflags -static -X github.com/chrislusf/seaweedfs/weed/util.COMMIT=${{github.sha}}
           # Where to run `go build .`
@@ -55,6 +56,7 @@ jobs:
           goarch: ${{ matrix.goarch }}
           release_tag: dev
           overwrite: true
+          pre_command: export CGO_ENABLED=0
           ldflags: -extldflags -static -X github.com/chrislusf/seaweedfs/weed/util.COMMIT=${{github.sha}}
           # Where to run `go build .`
           project_path: weed

README.md

@@ -7,6 +7,7 @@
 [![GoDoc](https://godoc.org/github.com/chrislusf/seaweedfs/weed?status.svg)](https://godoc.org/github.com/chrislusf/seaweedfs/weed)
 [![Wiki](https://img.shields.io/badge/docs-wiki-blue.svg)](https://github.com/chrislusf/seaweedfs/wiki)
 [![Docker Pulls](https://img.shields.io/docker/pulls/chrislusf/seaweedfs.svg?maxAge=4800)](https://hub.docker.com/r/chrislusf/seaweedfs/)
+[![SeaweedFS on Maven Central](https://img.shields.io/maven-central/v/com.github.chrislusf/seaweedfs-client)](https://search.maven.org/search?q=g:com.github.chrislusf)
 
 
 ![SeaweedFS Logo](https://raw.githubusercontent.com/chrislusf/seaweedfs/master/note/seaweedfs.png)
@@ -42,6 +43,7 @@ Your support will be really appreciated by me and other supporters!
 - [SeaweedFS on Twitter](https://twitter.com/SeaweedFS)
 - [SeaweedFS Mailing List](https://groups.google.com/d/forum/seaweedfs)
 - [Wiki Documentation](https://github.com/chrislusf/seaweedfs/wiki)
+- [SeaweedFS White Paper](https://github.com/chrislusf/seaweedfs/wiki/SeaweedFS_Architecture.pdf)
 - [SeaweedFS Introduction Slides](https://www.slideshare.net/chrislusf/seaweedfs-introduction)
 
 Table of Contents
@@ -70,7 +72,7 @@ Table of Contents
 * Download the latest binary from https://github.com/chrislusf/seaweedfs/releases and unzip a single binary file `weed` or `weed.exe`
 * Run `weed server -dir=/some/data/dir -s3` to start one master, one volume server, one filer, and one S3 gateway.
 
-Also, to increase capacity, just add more volume servers by running `weed volume -dir="/some/data/dir2" -mserver="<master_host>:9333" -port=8081` locally, or on a different machine, or on thoudsands of machines. That is it!
+Also, to increase capacity, just add more volume servers by running `weed volume -dir="/some/data/dir2" -mserver="<master_host>:9333" -port=8081` locally, or on a different machine, or on thousands of machines. That is it!
 
 ## Introduction ##
 
@@ -79,17 +81,34 @@ SeaweedFS is a simple and highly scalable distributed file system. There are two
 1. to store billions of files!
 2. to serve the files fast!
 
-SeaweedFS started as an Object Store to handle small files efficiently. Instead of managing all file metadata in a central master, the central master only manages volumes on volume servers, and these volume servers manage files and their metadata. This relieves concurrency pressure from the central master and spreads file metadata into volume servers, allowing faster file access (O(1), usually just one disk read operation).
+SeaweedFS started as an Object Store to handle small files efficiently. 
+Instead of managing all file metadata in a central master, 
+the central master only manages volumes on volume servers, 
+and these volume servers manage files and their metadata. 
+This relieves concurrency pressure from the central master and spreads file metadata into volume servers, 
+allowing faster file access (O(1), usually just one disk read operation).
 
-SeaweedFS can transparently integrate with the cloud. With hot data on local cluster, and warm data on the cloud with O(1) access time, SeaweedFS can achieve both fast local access time and elastic cloud storage capacity. What's more, the cloud storage access API cost is minimized. Faster and Cheaper than direct cloud storage!
+SeaweedFS can transparently integrate with the cloud. 
+With hot data on local cluster, and warm data on the cloud with O(1) access time, 
+SeaweedFS can achieve both fast local access time and elastic cloud storage capacity. 
+What's more, the cloud storage access API cost is minimized. 
+Faster and Cheaper than direct cloud storage!
+Signup for future managed SeaweedFS cluster offering at "seaweedfilesystem at gmail dot com".
 
-There is only 40 bytes of disk storage overhead for each file's metadata. It is so simple with O(1) disk reads that you are welcome to challenge the performance with your actual use cases.
+There is only 40 bytes of disk storage overhead for each file's metadata. 
+It is so simple with O(1) disk reads that you are welcome to challenge the performance with your actual use cases.
 
-SeaweedFS started by implementing [Facebook's Haystack design paper](http://www.usenix.org/event/osdi10/tech/full_papers/Beaver.pdf). Also, SeaweedFS implements erasure coding with ideas from [f4: Facebook’s Warm BLOB Storage System](https://www.usenix.org/system/files/conference/osdi14/osdi14-paper-muralidhar.pdf)
+SeaweedFS started by implementing [Facebook's Haystack design paper](http://www.usenix.org/event/osdi10/tech/full_papers/Beaver.pdf). 
+Also, SeaweedFS implements erasure coding with ideas from 
+[f4: Facebook’s Warm BLOB Storage System](https://www.usenix.org/system/files/conference/osdi14/osdi14-paper-muralidhar.pdf), and has a lot of similarities with [Facebook’s Tectonic Filesystem](https://www.usenix.org/system/files/fast21-pan.pdf)
 
-On top of the object store, optional [Filer] can support directories and POSIX attributes. Filer is a separate linearly-scalable stateless server with customizable metadata stores, e.g., MySql, Postgres, Redis, Cassandra, HBase, Mongodb, Elastic Search, LevelDB, RocksDB, MemSql, TiDB, Etcd, CockroachDB, etc.
+On top of the object store, optional [Filer] can support directories and POSIX attributes. 
+Filer is a separate linearly-scalable stateless server with customizable metadata stores, 
+e.g., MySql, Postgres, Redis, Cassandra, HBase, Mongodb, Elastic Search, LevelDB, RocksDB, MemSql, TiDB, Etcd, CockroachDB, etc.
 
-For any distributed key value stores, the large values can be offloaded to SeaweedFS. With the fast access speed and linearly scalable capacity, SeaweedFS can work as a distributed [Key-Large-Value store][KeyLargeValueStore].
+For any distributed key value stores, the large values can be offloaded to SeaweedFS. 
+With the fast access speed and linearly scalable capacity, 
+SeaweedFS can work as a distributed [Key-Large-Value store][KeyLargeValueStore].
 
 [Back to TOC](#table-of-contents)
 
@@ -105,6 +124,7 @@ For any distributed key value stores, the large values can be offloaded to Seawe
 * Support ETag, Accept-Range, Last-Modified, etc.
 * Support in-memory/leveldb/readonly mode tuning for memory/performance balance.
 * Support rebalancing the writable and readonly volumes.
+* [Customizable Multiple Storage Tiers][TieredStorage]: Customizable storage disk types to balance performance and cost.
 * [Transparent cloud integration][CloudTier]: unlimited capacity via tiered cloud storage for warm data.
 * [Erasure Coding for warm storage][ErasureCoding]  Rack-Aware 10.4 erasure coding reduces storage cost and increases availability.
 
@@ -135,6 +155,7 @@ For any distributed key value stores, the large values can be offloaded to Seawe
 [Hadoop]: https://github.com/chrislusf/seaweedfs/wiki/Hadoop-Compatible-File-System
 [WebDAV]: https://github.com/chrislusf/seaweedfs/wiki/WebDAV
 [ErasureCoding]: https://github.com/chrislusf/seaweedfs/wiki/Erasure-coding-for-warm-storage
+[TieredStorage]: https://github.com/chrislusf/seaweedfs/wiki/Tiered-Storage
 [CloudTier]: https://github.com/chrislusf/seaweedfs/wiki/Cloud-Tier
 [FilerDataEncryption]: https://github.com/chrislusf/seaweedfs/wiki/Filer-Data-Encryption
 [FilerTTL]: https://github.com/chrislusf/seaweedfs/wiki/Filer-Stores

backers.md

@@ -1,3 +1,4 @@
+
 <h1 align="center">Sponsors &amp; Backers</h1>
 
 - [Become a backer or sponsor on Patreon](https://www.patreon.com/seaweedfs).
@@ -12,3 +13,4 @@
 
 - [ColorfulClouds Tech Co. Ltd.](https://caiyunai.com/)
 - [Haravan - Ecommerce Platform](https://www.haravan.com)
+- PeterCxy - Creator of Shelter App

docker/Dockerfile.go_build

@@ -6,7 +6,7 @@ ARG BRANCH=${BRANCH:-master}
 RUN cd /go/src/github.com/chrislusf/seaweedfs && git checkout $BRANCH
 RUN cd /go/src/github.com/chrislusf/seaweedfs/weed \
   && export LDFLAGS="-X github.com/chrislusf/seaweedfs/weed/util.COMMIT=$(git rev-parse --short HEAD)" \
-  && go install -ldflags "${LDFLAGS}"
+  && CGO_ENABLED=0 go install -ldflags "-extldflags -static ${LDFLAGS}"
 
 FROM alpine AS final
 LABEL author="Chris Lu"

docker/Dockerfile.go_build_large

@@ -6,7 +6,7 @@ ARG BRANCH=${BRANCH:-master}
 RUN cd /go/src/github.com/chrislusf/seaweedfs && git checkout $BRANCH
 RUN cd /go/src/github.com/chrislusf/seaweedfs/weed \
   && export LDFLAGS="-X github.com/chrislusf/seaweedfs/weed/util.COMMIT=$(git rev-parse --short HEAD)" \
-  && go install -tags 5BytesOffset -ldflags "${LDFLAGS}"
+  && CGO_ENABLED=0 go install -tags 5BytesOffset -ldflags "-extldflags -static ${LDFLAGS}"
 
 FROM alpine AS final
 LABEL author="Chris Lu"

docker/Dockerfile.s3tests

@@ -28,4 +28,4 @@ ENV \
     S3TEST_CONF="/s3test.conf"
 
 ENTRYPOINT ["/bin/bash", "-c"]
-CMD ["exec ./virtualenv/bin/nosetests ${NOSETESTS_OPTIONS-} ${NOSETESTS_ATTR:+-a $NOSETESTS_ATTR} ${NOSETESTS_EXCLUDE:+-e $NOSETESTS_EXCLUDE}"]
+CMD ["sleep 10 && exec ./virtualenv/bin/nosetests ${NOSETESTS_OPTIONS-} ${NOSETESTS_ATTR:+-a $NOSETESTS_ATTR} ${NOSETESTS_EXCLUDE:+-e $NOSETESTS_EXCLUDE}"]

docker/Makefile

@@ -4,8 +4,10 @@ all: gen
 
 gen: dev
 
-build:
-	cd ../weed; GOOS=linux go build; mv weed ../docker/
+binary:
+	cd ../weed; CGO_ENABLED=0 GOOS=linux go build -ldflags "-extldflags -static"; mv weed ../docker/
+
+build: binary
 	docker build --no-cache -t chrislusf/seaweedfs:local -f Dockerfile.local .
 	rm ./weed
 
@@ -15,9 +17,15 @@ s3tests_build:
 dev: build
 	docker-compose -f compose/local-dev-compose.yml -p seaweedfs up
 
+dev_tls: build certstrap
+	ENV_FILE="tls.env" docker-compose -f compose/local-dev-compose.yml -p seaweedfs up
+
 dev_mount: build
 	docker-compose -f compose/local-mount-compose.yml -p seaweedfs up
 
+profile_mount: build
+	docker-compose -f compose/local-mount-profile-compose.yml -p seaweedfs up
+
 k8s: build
 	docker-compose -f compose/local-k8s-compose.yml -p seaweedfs up
 
@@ -41,3 +49,15 @@ filer_etcd: build
 
 clean:
 	rm ./weed
+
+certstrap:
+	go get github.com/square/certstrap
+	certstrap --depot-path compose/tls init --passphrase "" --common-name "SeaweedFS CA" || true
+	certstrap --depot-path compose/tls request-cert --passphrase "" --common-name volume01.dev || true
+	certstrap --depot-path compose/tls request-cert --passphrase "" --common-name master01.dev || true
+	certstrap --depot-path compose/tls request-cert --passphrase "" --common-name filer01.dev || true
+	certstrap --depot-path compose/tls request-cert --passphrase "" --common-name client01.dev || true
+	certstrap --depot-path compose/tls sign --CA "SeaweedFS CA" volume01.dev || true
+	certstrap --depot-path compose/tls sign --CA "SeaweedFS CA" master01.dev || true
+	certstrap --depot-path compose/tls sign --CA "SeaweedFS CA" filer01.dev || true
+	certstrap --depot-path compose/tls sign --CA "SeaweedFS CA" client01.dev || true

docker/compose/local-clusters-compose.yml

@@ -11,6 +11,10 @@ services:
       - 8888:8888
       - 18888:18888
     command: "server -ip=server1 -filer -volume.max=0 -master.volumeSizeLimitMB=1024 -volume.preStopSeconds=1"
+    volumes:
+      - ./master-cloud.toml:/etc/seaweedfs/master.toml
+    depends_on:
+      - server2
   server2:
     image: chrislusf/seaweedfs:local
     ports:
@@ -20,4 +24,5 @@ services:
       - 18085:18080
       - 8889:8888
       - 18889:18888
-    command: "server -ip=server2 -filer -volume.max=0 -master.volumeSizeLimitMB=1024 -volume.preStopSeconds=1"
+      - 8334:8333
+    command: "server -ip=server2 -filer -s3 -volume.max=0 -master.volumeSizeLimitMB=1024 -volume.preStopSeconds=1"

docker/compose/local-dev-compose.yml

@@ -6,33 +6,49 @@ services:
     ports:
       - 9333:9333
       - 19333:19333
-    command: "master -ip=master"
+    command: "-v=1 master -ip=master"
+    volumes:
+      - ./tls:/etc/seaweedfs/tls
+    env_file:
+     - ${ENV_FILE:-dev.env}
   volume:
     image: chrislusf/seaweedfs:local
     ports:
       - 8080:8080
       - 18080:18080
-    command: "volume -mserver=master:9333 -port=8080 -ip=volume -preStopSeconds=1"
+    command: "-v=1 volume -mserver=master:9333 -port=8080 -ip=volume -preStopSeconds=1"
     depends_on:
       - master
+    volumes:
+      - ./tls:/etc/seaweedfs/tls
+    env_file:
+      - ${ENV_FILE:-dev.env}
   filer:
     image: chrislusf/seaweedfs:local
     ports:
       - 8888:8888
       - 18888:18888
-    command: 'filer -master="master:9333"'
+    command: '-v=1 filer -master="master:9333"'
     depends_on:
       - master
       - volume
+    volumes:
+      - ./tls:/etc/seaweedfs/tls
+    env_file:
+      - ${ENV_FILE:-dev.env}
   s3:
     image: chrislusf/seaweedfs:local
     ports:
       - 8333:8333
-    command: 's3 -filer="filer:8888"'
+    command: '-v=1 s3 -filer="filer:8888"'
     depends_on:
       - master
       - volume
       - filer
+    volumes:
+      - ./tls:/etc/seaweedfs/tls
+    env_file:
+      - ${ENV_FILE:-dev.env}
   mount:
     image: chrislusf/seaweedfs:local
     privileged: true
@@ -40,6 +56,10 @@ services:
       - SYS_ADMIN
     mem_limit: 4096m
     command: '-v=4 mount -filer="filer:8888" -dirAutoCreate -dir=/mnt/seaweedfs -cacheCapacityMB=100 -concurrentWriters=128'
+    volumes:
+      - ./tls:/etc/seaweedfs/tls
+    env_file:
+      - ${ENV_FILE:-dev.env}
     depends_on:
       - master
       - volume

docker/compose/local-minio-gateway-compose.yml

@@ -0,0 +1,50 @@
+version: '2'
+
+services:
+  master:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 9333:9333
+      - 19333:19333
+    command: "master -ip=master -volumeSizeLimitMB=1024"
+  volume:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8080:8080
+      - 18080:18080
+    command: "volume -mserver=master:9333 -port=8080 -ip=volume -max=0 -preStopSeconds=1"
+    depends_on:
+      - master
+  s3:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8888:8888
+      - 18888:18888
+      - 8333:8333
+    command: '-v 1 filer -master="master:9333" -s3 -s3.config=/etc/seaweedfs/s3.json -s3.port=8333'
+    volumes:
+      - ./s3.json:/etc/seaweedfs/s3.json
+    depends_on:
+      - master
+      - volume
+  minio-gateway-s3:
+    image: minio/minio
+    ports:
+      - 9000:9000
+    command: 'minio gateway s3 http://s3:8333'
+    restart: on-failure
+    environment:
+      MINIO_ACCESS_KEY: "some_access_key1"
+      MINIO_SECRET_KEY: "some_secret_key1"
+    depends_on:
+      - s3
+  minio-warp:
+    image: minio/warp
+    command: 'mixed --duration=5m --obj.size=3mb --autoterm'
+    restart: on-failure
+    environment:
+      WARP_HOST: "minio-gateway-s3:9000"
+      WARP_ACCESS_KEY: "some_access_key1"
+      WARP_SECRET_KEY: "some_secret_key1"
+    depends_on:
+      - minio-gateway-s3

docker/compose/local-mount-profile-compose.yml

@@ -0,0 +1,47 @@
+version: '2'
+
+services:
+  master:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 9333:9333
+      - 19333:19333
+    command: "master -ip=master"
+  volume:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 7455:8080
+      - 9325:9325
+    volumes:
+      - /Volumes/mobile_disk/99:/data
+    command: 'volume -mserver="master:9333" -port=8080 -metricsPort=9325 -preStopSeconds=1 -publicUrl=localhost:7455'
+    depends_on:
+      - master
+  filer:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8888:8888
+      - 18888:18888
+      - 9326:9326
+    volumes:
+      - /Volumes/mobile_disk/99:/data
+    command: 'filer -master="master:9333"  -metricsPort=9326'
+    tty: true
+    stdin_open: true
+    depends_on:
+      - master
+      - volume
+  mount:
+    image: chrislusf/seaweedfs:local
+    privileged: true
+    cap_add:
+      - SYS_ADMIN
+    devices:
+      - fuse
+    volumes:
+      - /Volumes/mobile_disk/99:/data
+    entrypoint: '/bin/sh -c "mkdir -p t1 && weed mount -filer=filer:8888 -dir=./t1 -cacheCapacityMB=0 -memprofile=/data/mount.mem.pprof"'
+    depends_on:
+      - master
+      - volume
+      - filer

docker/compose/local-registry-compose.yml

@@ -15,24 +15,18 @@ services:
     command: "volume -mserver=master:9333 -port=8080 -ip=volume -max=0 -preStopSeconds=1"
     depends_on:
       - master
-  filer:
+  s3:
     image: chrislusf/seaweedfs:local
     ports:
       - 8888:8888
       - 18888:18888
-    command: 'filer -master="master:9333"'
-    depends_on:
-      - master
-      - volume
-  s3:
-    image: chrislusf/seaweedfs:local
-    ports:
       - 8333:8333
-    command: '-v 9 s3 -filer="filer:8888"'
+    command: '-v 9 filer -master="master:9333" -s3 -s3.config=/etc/seaweedfs/s3.json -s3.port=8333'
+    volumes:
+      - ./s3.json:/etc/seaweedfs/s3.json
     depends_on:
       - master
       - volume
-      - filer
   minio:
     image: minio/minio
     ports:

docker/compose/local-s3tests-compose.yml

@@ -38,7 +38,7 @@ services:
       S3TEST_CONF: "s3tests.conf"
       NOSETESTS_OPTIONS: "--verbose --logging-level=ERROR --with-xunit --failure-detail s3tests_boto3.functional.test_s3"
       NOSETESTS_ATTR: "!tagging,!fails_on_aws,!encryption,!bucket-policy,!versioning,!fails_on_rgw,!bucket-policy,!fails_with_subdomain,!policy_status,!object-lock,!lifecycle,!cors,!user-policy"
-      NOSETESTS_EXCLUDE: "(bucket_list_delimiter_basic|bucket_listv2_delimiter_basic|bucket_listv2_encoding_basic|bucket_list_encoding_basic|bucket_list_delimiter_prefix|bucket_listv2_delimiter_prefix_ends_with_delimiter|bucket_list_delimiter_prefix_ends_with_delimiter|bucket_list_delimiter_alt|bucket_listv2_delimiter_alt|bucket_list_delimiter_prefix_underscore|bucket_list_delimiter_percentage|bucket_listv2_delimiter_percentage|bucket_list_delimiter_whitespace|bucket_listv2_delimiter_whitespace|bucket_list_delimiter_dot|bucket_listv2_delimiter_dot|bucket_list_delimiter_unreadable|bucket_listv2_delimiter_unreadable|bucket_listv2_fetchowner_defaultempty|bucket_listv2_fetchowner_empty|bucket_list_delimiter_not_skip_special|bucket_list_prefix_delimiter_alt|bucket_listv2_prefix_delimiter_alt|bucket_list_prefix_delimiter_prefix_not_exist|bucket_listv2_prefix_delimiter_prefix_not_exist|bucket_list_prefix_delimiter_delimiter_not_exist|bucket_listv2_prefix_delimiter_delimiter_not_exist|bucket_list_prefix_delimiter_prefix_delimiter_not_exist|bucket_listv2_prefix_delimiter_prefix_delimiter_not_exist|bucket_list_maxkeys_none|bucket_listv2_maxkeys_none|bucket_list_maxkeys_invalid|bucket_listv2_continuationtoken_empty|bucket_list_return_data|bucket_list_objects_anonymous|bucket_listv2_objects_anonymous|bucket_notexist|bucketv2_notexist|bucket_delete_nonempty|bucket_concurrent_set_canned_acl|object_write_to_nonexist_bucket|object_requestid_matches_header_on_error|object_head_zero_bytes|object_write_cache_control|object_write_expires|object_set_get_metadata_none_to_good|object_set_get_metadata_none_to_empty|object_set_get_metadata_overwrite_to_empty|post_object_anonymous_request|post_object_authenticated_request|post_object_authenticated_no_content_type|post_object_authenticated_request_bad_access_key|post_object_set_success_code|post_object_set_invalid_success_code|post_object_upload_larger_than_chunk|post_object_set_key_from_filename|post_object_ignored_header|post_object_case_insensitive_condition_fields|post_object_escaped_field_values|post_object_success_redirect_action|post_object_invalid_signature|post_object_invalid_access_key|post_object_missing_policy_condition|post_object_user_specified_header|post_object_request_missing_policy_specified_field|post_object_expired_policy|post_object_invalid_request_field_value|get_object_ifmatch_failed|get_object_ifunmodifiedsince_good|put_object_ifmatch_failed|object_raw_get|object_raw_get_bucket_gone|object_delete_key_bucket_gone|object_raw_get_bucket_acl|object_raw_get_object_acl|object_raw_authenticated|object_raw_response_headers|object_raw_authenticated_bucket_acl|object_raw_authenticated_object_acl|object_raw_authenticated_bucket_gone|object_raw_get_x_amz_expires_not_expired|object_raw_get_x_amz_expires_out_max_range|object_raw_get_x_amz_expires_out_positive_range|object_anon_put_write_access|object_raw_put_authenticated_expired|bucket_create_naming_bad_short_one|bucket_create_naming_bad_short_two|bucket_create_exists|bucket_get_location|bucket_acl_default|bucket_acl_canned|bucket_acl_canned_publicreadwrite|bucket_acl_canned_authenticatedread|object_acl_default|object_acl_canned_during_create|object_acl_canned|object_acl_canned_publicreadwrite|object_acl_canned_authenticatedread|object_acl_canned_bucketownerread|object_acl_canned_bucketownerfullcontrol|object_acl_full_control_verify_attributes|bucket_acl_canned_private_to_private|bucket_acl_grant_nonexist_user|bucket_acl_no_grants|bucket_acl_grant_email_not_exist|bucket_acl_revoke_all|bucket_recreate_not_overriding|bucket_create_special_key_names|object_copy_zero_size|object_copy_verify_contenttype|object_copy_to_itself|object_copy_to_itself_with_metadata|object_copy_not_owned_bucket|object_copy_not_owned_object_bucket|object_copy_retaining_metadata|object_copy_replacing_metadata|multipart_upload_empty|multipart_copy_invalid_range|multipart_copy_special_names|multipart_upload_resend_part|multipart_upload_size_too_small|abort_multipart_upload_not_found|multipart_upload_missing_part|multipart_upload_incorrect_etag|100_continue|ranged_request_invalid_range|ranged_request_empty_object|access_bucket)"
+      NOSETESTS_EXCLUDE: "(bucket_list_delimiter_basic|bucket_listv2_delimiter_basic|bucket_listv2_encoding_basic|bucket_list_encoding_basic|bucket_list_delimiter_prefix|bucket_listv2_delimiter_prefix_ends_with_delimiter|bucket_list_delimiter_prefix_ends_with_delimiter|bucket_list_delimiter_alt|bucket_listv2_delimiter_alt|bucket_list_delimiter_prefix_underscore|bucket_list_delimiter_percentage|bucket_listv2_delimiter_percentage|bucket_list_delimiter_whitespace|bucket_listv2_delimiter_whitespace|bucket_list_delimiter_dot|bucket_listv2_delimiter_dot|bucket_list_delimiter_unreadable|bucket_listv2_delimiter_unreadable|bucket_listv2_fetchowner_defaultempty|bucket_listv2_fetchowner_empty|bucket_list_prefix_delimiter_alt|bucket_listv2_prefix_delimiter_alt|bucket_list_prefix_delimiter_prefix_not_exist|bucket_listv2_prefix_delimiter_prefix_not_exist|bucket_list_prefix_delimiter_delimiter_not_exist|bucket_listv2_prefix_delimiter_delimiter_not_exist|bucket_list_prefix_delimiter_prefix_delimiter_not_exist|bucket_listv2_prefix_delimiter_prefix_delimiter_not_exist|bucket_list_maxkeys_none|bucket_listv2_maxkeys_none|bucket_list_maxkeys_invalid|bucket_listv2_continuationtoken_empty|bucket_list_return_data|bucket_list_objects_anonymous|bucket_listv2_objects_anonymous|bucket_notexist|bucketv2_notexist|bucket_delete_nonempty|bucket_concurrent_set_canned_acl|object_write_to_nonexist_bucket|object_requestid_matches_header_on_error|object_write_cache_control|object_write_expires|object_set_get_metadata_none_to_good|object_set_get_metadata_none_to_empty|object_set_get_metadata_overwrite_to_empty|post_object_anonymous_request|post_object_authenticated_request|post_object_authenticated_no_content_type|post_object_authenticated_request_bad_access_key|post_object_set_success_code|post_object_set_invalid_success_code|post_object_upload_larger_than_chunk|post_object_set_key_from_filename|post_object_ignored_header|post_object_case_insensitive_condition_fields|post_object_escaped_field_values|post_object_success_redirect_action|post_object_invalid_signature|post_object_invalid_access_key|post_object_missing_policy_condition|post_object_user_specified_header|post_object_request_missing_policy_specified_field|post_object_expired_policy|post_object_invalid_request_field_value|get_object_ifmatch_failed|get_object_ifunmodifiedsince_good|put_object_ifmatch_failed|object_raw_get_bucket_gone|object_delete_key_bucket_gone|object_raw_get_bucket_acl|object_raw_get_object_acl|object_raw_response_headers|object_raw_authenticated_bucket_gone|object_raw_get_x_amz_expires_out_max_range|object_raw_get_x_amz_expires_out_positive_range|object_anon_put_write_access|object_raw_put_authenticated_expired|bucket_create_exists|bucket_create_naming_bad_short_one|bucket_create_naming_bad_short_two|bucket_get_location|bucket_acl_default|bucket_acl_canned|bucket_acl_canned_publicreadwrite|bucket_acl_canned_authenticatedread|object_acl_default|object_acl_canned_during_create|object_acl_canned|object_acl_canned_publicreadwrite|object_acl_canned_authenticatedread|object_acl_canned_bucketownerread|object_acl_canned_bucketownerfullcontrol|object_acl_full_control_verify_attributes|bucket_acl_canned_private_to_private|bucket_acl_grant_nonexist_user|bucket_acl_no_grants|bucket_acl_grant_email_not_exist|bucket_acl_revoke_all|bucket_recreate_not_overriding|object_copy_verify_contenttype|object_copy_to_itself_with_metadata|object_copy_not_owned_bucket|object_copy_not_owned_object_bucket|object_copy_retaining_metadata|object_copy_replacing_metadata|multipart_upload_empty|multipart_copy_invalid_range|multipart_copy_special_names|multipart_upload_resend_part|multipart_upload_size_too_small|abort_multipart_upload_not_found|multipart_upload_missing_part|multipart_upload_incorrect_etag|100_continue|ranged_request_invalid_range|ranged_request_empty_object|access_bucket)"
     depends_on:
       - master
       - volume

docker/compose/master-cloud.toml

@@ -0,0 +1,30 @@
+
+# Put this file to one of the location, with descending priority
+#    ./master.toml
+#    $HOME/.seaweedfs/master.toml
+#    /etc/seaweedfs/master.toml
+# this file is read by master
+
+[master.maintenance]
+# periodically run these scripts are the same as running them from 'weed shell'
+scripts = """
+  lock
+  ec.encode -fullPercent=95 -quietFor=1h
+  ec.rebuild -force
+  ec.balance -force
+  volume.balance -force
+  volume.fix.replication
+  unlock
+"""
+sleep_minutes = 17          # sleep minutes between each script execution
+
+# configurations for tiered cloud storage
+# old volumes are transparently moved to cloud for cost efficiency
+[storage.backend]
+	[storage.backend.s3.default]
+	enabled = true
+	aws_access_key_id     = "any"     # if empty, loads from the shared credentials file (~/.aws/credentials).
+	aws_secret_access_key = "any"     # if empty, loads from the shared credentials file (~/.aws/credentials).
+	region = "us-east-2"
+	bucket = "volume_bucket"    # an existing bucket
+	endpoint = "http://server2:8333"

docker/compose/tls.env

@@ -0,0 +1,14 @@
+WEED_GRPC_CA=/etc/seaweedfs/tls/SeaweedFS_CA.crt
+WEED_GRPC_ALLOWED_WILDCARD_DOMAIN=".dev"
+WEED_GRPC_MASTER_CERT=/etc/seaweedfs/tls/master01.dev.crt
+WEED_GRPC_MASTER_KEY=/etc/seaweedfs/tls/master01.dev.key
+WEED_GRPC_VOLUME_CERT=/etc/seaweedfs/tls/volume01.dev.crt
+WEED_GRPC_VOLUME_KEY=/etc/seaweedfs/tls/volume01.dev.key
+WEED_GRPC_FILER_CERT=/etc/seaweedfs/tls/filer01.dev.crt
+WEED_GRPC_FILER_KEY=/etc/seaweedfs/tls/filer01.dev.key
+WEED_GRPC_CLIENT_CERT=/etc/seaweedfs/tls/client01.dev.crt
+WEED_GRPC_CLIENT_KEY=/etc/seaweedfs/tls/client01.dev.key
+WEED_GRPC_MASTER_ALLOWED_COMMONNAMES="volume01.dev,master01.dev,filer01.dev,client01.dev"
+WEED_GRPC_VOLUME_ALLOWED_COMMONNAMES="volume01.dev,master01.dev,filer01.dev,client01.dev"
+WEED_GRPC_FILER_ALLOWED_COMMONNAMES="volume01.dev,master01.dev,filer01.dev,client01.dev"
+WEED_GRPC_CLIENT_ALLOWED_COMMONNAMES="volume01.dev,master01.dev,filer01.dev,client01.dev"

docker/entrypoint.sh

@@ -60,9 +60,9 @@ case "$1" in
   'cronjob')
 	MASTER=${WEED_MASTER-localhost:9333}
 	FIX_REPLICATION_CRON_SCHEDULE=${CRON_SCHEDULE-*/7 * * * * *}
-	echo "$FIX_REPLICATION_CRON_SCHEDULE" 'echo "volume.fix.replication" | weed shell -master='$MASTER > /crontab
+	echo "$FIX_REPLICATION_CRON_SCHEDULE" 'echo "lock; volume.fix.replication; unlock" | weed shell -master='$MASTER > /crontab
 	BALANCING_CRON_SCHEDULE=${CRON_SCHEDULE-25 * * * * *}
-	echo "$BALANCING_CRON_SCHEDULE" 'echo "volume.balance -c ALL -force" | weed shell -master='$MASTER >> /crontab
+	echo "$BALANCING_CRON_SCHEDULE" 'echo "lock; volume.balance -collection ALL_COLLECTIONS -force; unlock" | weed shell -master='$MASTER >> /crontab
 	echo "Running Crontab:"
 	cat /crontab
 	exec supercronic /crontab

go.mod

@@ -13,6 +13,7 @@ require (
 	github.com/Shopify/sarama v1.23.1
 	github.com/aws/aws-sdk-go v1.34.30
 	github.com/buraksezer/consistent v0.0.0-20191006190839-693edf70fd72
+	github.com/bwmarrin/snowflake v0.3.0
 	github.com/cespare/xxhash v1.1.0
 	github.com/chrislusf/raft v1.0.4
 	github.com/coreos/go-semver v0.3.0 // indirect
@@ -38,6 +39,7 @@ require (
 	github.com/google/uuid v1.1.1
 	github.com/gorilla/mux v1.7.4
 	github.com/gorilla/websocket v1.4.1 // indirect
+	github.com/grpc-ecosystem/go-grpc-middleware v1.0.1-0.20190118093823-f849b5445de4
 	github.com/grpc-ecosystem/grpc-gateway v1.11.0 // indirect
 	github.com/jcmturner/gofork v1.0.0 // indirect
 	github.com/json-iterator/go v1.1.10
@@ -48,7 +50,7 @@ require (
 	github.com/klauspost/crc32 v1.2.0
 	github.com/klauspost/reedsolomon v1.9.2
 	github.com/kurin/blazer v0.5.3
-	github.com/lib/pq v1.2.0
+	github.com/lib/pq v1.10.0
 	github.com/lunixbochs/vtclean v1.0.0 // indirect
 	github.com/magiconair/properties v1.8.1 // indirect
 	github.com/mattn/go-colorable v0.1.2 // indirect
@@ -59,7 +61,7 @@ require (
 	github.com/prometheus/client_golang v1.3.0
 	github.com/rakyll/statik v0.1.7
 	github.com/rcrowley/go-metrics v0.0.0-20190826022208-cac0b30c2563 // indirect
-	github.com/seaweedfs/fuse v1.1.1
+	github.com/seaweedfs/fuse v1.1.3
 	github.com/seaweedfs/goexif v1.0.2
 	github.com/skip2/go-qrcode v0.0.0-20200617195104-da1b6568686e
 	github.com/spaolacci/murmur3 v1.1.0 // indirect
@@ -87,7 +89,7 @@ require (
 	gocloud.dev/pubsub/rabbitpubsub v0.20.0
 	golang.org/x/image v0.0.0-20200119044424-58c23975cae1 // indirect
 	golang.org/x/net v0.0.0-20201202161906-c7110b5ffcbb
-	golang.org/x/sync v0.0.0-20200930132711-30421366ff76 // indirect
+	golang.org/x/sync v0.0.0-20210220032951-036812b2e83c
 	golang.org/x/sys v0.0.0-20201022201747-fb209a7c41cd
 	golang.org/x/tools v0.0.0-20200608174601-1b747fd94509
 	google.golang.org/api v0.26.0

go.sum

@@ -141,6 +141,8 @@ github.com/bmizerany/assert v0.0.0-20160611221934-b7ed37b82869 h1:DDGfHa7BWjL4Yn
 github.com/bmizerany/assert v0.0.0-20160611221934-b7ed37b82869/go.mod h1:Ekp36dRnpXw/yCqJaO+ZrUyxD+3VXMFFr56k5XYrpB4=
 github.com/buraksezer/consistent v0.0.0-20191006190839-693edf70fd72 h1:fUmDBbSvv1uOzo/t8WaxZMVb7BxJ8JECo5lGoR9c5bA=
 github.com/buraksezer/consistent v0.0.0-20191006190839-693edf70fd72/go.mod h1:OEE5igu/CDjGegM1Jn6ZMo7R6LlV/JChAkjfQQIRLpg=
+github.com/bwmarrin/snowflake v0.3.0 h1:xm67bEhkKh6ij1790JB83OujPR5CzNe8QuQqAgISZN0=
+github.com/bwmarrin/snowflake v0.3.0/go.mod h1:NdZxfVWX+oR6y2K0o6qAYv6gIOP9rjG0/E9WsDpxqwE=
 github.com/casbin/casbin/v2 v2.1.2/go.mod h1:YcPU1XXisHhLzuxH9coDNf2FbKpjGlbCg3n9yuLkIJQ=
 github.com/cenkalti/backoff v2.2.1+incompatible/go.mod h1:90ReRw6GdpyfrHakVjL/QHaoyV4aDUVVkXQJJJ3NXXM=
 github.com/census-instrumentation/opencensus-proto v0.2.0 h1:LzQXZOgg4CQfE6bFvXGM30YZL1WW/M337pXml+GrcZ4=
@@ -496,6 +498,8 @@ github.com/kurin/blazer v0.5.3/go.mod h1:4FCXMUWo9DllR2Do4TtBd377ezyAJ51vB5uTBjt
 github.com/lib/pq v1.1.1/go.mod h1:5WUZQaWbwv1U+lTReE5YruASi9Al49XbQIvNi/34Woo=
 github.com/lib/pq v1.2.0 h1:LXpIM/LZ5xGFhOpXAQUIMM1HdyqzVYM13zNdjCEEcA0=
 github.com/lib/pq v1.2.0/go.mod h1:5WUZQaWbwv1U+lTReE5YruASi9Al49XbQIvNi/34Woo=
+github.com/lib/pq v1.10.0 h1:Zx5DJFEYQXio93kgXnQ09fXNiUKsqv4OUEu2UtGcB1E=
+github.com/lib/pq v1.10.0/go.mod h1:AlVN5x4E4T544tWzH6hKfbfQvm3HdbOxrmggDNAPY9o=
 github.com/lightstep/lightstep-tracer-common/golang/gogo v0.0.0-20190605223551-bc2310a04743/go.mod h1:qklhhLq1aX+mtWk9cPHPzaBjWImj5ULL6C7HFJtXQMM=
 github.com/lightstep/lightstep-tracer-go v0.18.1/go.mod h1:jlF1pusYV4pidLvZ+XD0UBX0ZE6WURAspgAczcDHrL4=
 github.com/lunixbochs/vtclean v1.0.0 h1:xu2sLAri4lGiovBDQKxl5mrXyESr3gUr5m5SM5+LVb8=
@@ -681,6 +685,8 @@ github.com/seaweedfs/fuse v1.1.0 h1:cL1qPHFNtFv0UuJTLjKKgWDzfJ4iZzTa4Y7ipc2acGw=
 github.com/seaweedfs/fuse v1.1.0/go.mod h1:+PP6WlkrRUG6KPE+Th2EX5To/PjHaFsvqg/UgQ39aj8=
 github.com/seaweedfs/fuse v1.1.1 h1:WD51YFJcBViOx8I89jeqPD+vAKl4EowzBy9GUw0plb0=
 github.com/seaweedfs/fuse v1.1.1/go.mod h1:+PP6WlkrRUG6KPE+Th2EX5To/PjHaFsvqg/UgQ39aj8=
+github.com/seaweedfs/fuse v1.1.3 h1:0DddotXwSRGbYG2kynoJyr8GHCy30Z2SpdhP3vdyijY=
+github.com/seaweedfs/fuse v1.1.3/go.mod h1:+PP6WlkrRUG6KPE+Th2EX5To/PjHaFsvqg/UgQ39aj8=
 github.com/seaweedfs/goexif v1.0.2 h1:p+rTXYdQ2mgxd+1JaTrQ9N8DvYuw9UH9xgYmJ+Bb29E=
 github.com/seaweedfs/goexif v1.0.2/go.mod h1:MrKs5LK0HXdffrdCZrW3OIMegL2xXpC6ThLyXMyjdrk=
 github.com/secsy/goftp v0.0.0-20190720192957-f31499d7c79a h1:C6IhVTxNkhlb0tlCB6JfHOUv1f0xHPK7V8X4HlJZEJw=
@@ -955,6 +961,8 @@ golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e/go.mod h1:RxMgew5VJxzue5/jJ
 golang.org/x/sync v0.0.0-20200317015054-43a5402ce75a/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20200930132711-30421366ff76 h1:JnxiSYT3Nm0BT2a8CyvYyM6cnrWpidecD1UuSYbhKm0=
 golang.org/x/sync v0.0.0-20200930132711-30421366ff76/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20210220032951-036812b2e83c h1:5KslGYwFpkhGh+Q16bwMP3cOontH8FOep7tGV86Y7SQ=
+golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sys v0.0.0-20180823144017-11551d06cbcc/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=

k8s/README.md

@@ -29,6 +29,14 @@ please set/update the corresponding affinity rule in values.yaml to an empty one
 
 ```affinity: ""```
 
+### PVC - storage class ###
+
+on the volume stateful set added support for K8S PVC, currently example
+with the simple local-path-provisioner from Rancher (comes included with k3d / k3s)
+https://github.com/rancher/local-path-provisioner
+
+you can use ANY storage class you like, just update the correct storage-class
+for your deployment.
 
 ### current instances config (AIO):
 1 instance for each type (master/filer+s3/volume)

k8s/seaweedfs/Chart.yaml

@@ -1,5 +1,5 @@
 apiVersion: v1
 description: SeaweedFS
 name: seaweedfs
-appVersion: "2.27"
-version: 2.27
+appVersion: "2.38"
+version: 2.38

k8s/seaweedfs/templates/_helpers.tpl

@@ -126,3 +126,26 @@ Inject extra environment vars in the format key:value, if populated
 {{- printf "%s%s%s:%s" $registryName $repositoryName $name $tag -}}
 {{- end -}}
 {{- end -}}
+
+
+{{/* check if any PVC exists */}}
+{{- define "volume.pvc_exists" -}}
+{{- if or (or (eq .Values.volume.data.type "persistentVolumeClaim") (and (eq .Values.volume.idx.type "persistentVolumeClaim") .Values.volume.dir_idx )) (eq .Values.volume.logs.type "persistentVolumeClaim") -}}
+{{- printf "true" -}}
+{{- else -}}
+{{- printf "false" -}}
+{{- end -}}
+{{- end -}}
+
+{{/* check if any HostPath exists */}}
+{{- define "volume.hostpath_exists" -}}
+{{- if or (or (eq .Values.volume.data.type "hostPath") (and (eq .Values.volume.idx.type "hostPath") .Values.volume.dir_idx )) (eq .Values.volume.logs.type "hostPath") -}}
+{{- printf "true" -}}
+{{- else -}}
+{{- if or .Values.global.enableSecurity .Values.volume.extraVolumes -}}
+{{- printf "true" -}}
+{{- else -}}
+{{- printf "false" -}}
+{{- end -}}
+{{- end -}}
+{{- end -}}

k8s/seaweedfs/templates/cronjob.yaml

@@ -40,7 +40,7 @@ spec:
                   {{ if .Values.volume.dataCenter }} -dataCenter {{ .Values.volume.dataCenter }}{{ end }}\
                   {{ if .Values.cronjob.collection }} -collection {{ .Values.cronjob.collection }}{{ end }}\n\
                   {{- if .Values.cronjob.enableFixReplication }}
-                  volume.fix.replication {{ if .Values.cronjob.collectionPattern }} -collectionPattern={{ .Values.cronjob.collectionPattern }} {{ end }} \n\
+                  volume.fix.replication -collectionPattern={{ .Values.cronjob.collectionPattern }} \n\
                   {{- end }}
                   unlock\n" | \
                   /usr/bin/weed shell \

k8s/seaweedfs/templates/filer-service-client.yaml

@@ -10,6 +10,7 @@ metadata:
     monitoring: "true"
 {{- end }}
 spec:
+  clusterIP: None
   ports:
   - name: "swfs-filer"
     port: {{ .Values.filer.port }}

k8s/seaweedfs/templates/volume-statefulset.yaml

@@ -45,6 +45,19 @@ spec:
       priorityClassName: {{ .Values.volume.priorityClassName | quote }}
       {{- end }}
       enableServiceLinks: false
+      {{- if .Values.volume.dir_idx }}
+      initContainers:
+        - name: seaweedfs-vol-move-idx
+          image: {{ template "volume.image" . }}
+          imagePullPolicy: {{ .Values.global.pullPolicy | default "IfNotPresent" }}
+          command: [ '/bin/sh', '-c' ]
+          args: ['if ls {{ .Values.volume.dir }}/*.idx >/dev/null 2>&1; then mv {{ .Values.volume.dir }}/*.idx {{ .Values.volume.dir_idx }}/; fi;']
+          volumeMounts:
+          - name: idx
+            mountPath: {{ .Values.volume.dir_idx }}
+          - name: data
+            mountPath: {{ .Values.volume.dir }}
+      {{- end }}
       containers:
         - name: seaweedfs
           image: {{ template "volume.image" . }}
@@ -118,9 +131,13 @@ spec:
                 -compactionMBps={{ .Values.volume.compactionMBps }} \
                 -mserver={{ range $index := until (.Values.master.replicas | int) }}${SEAWEEDFS_FULLNAME}-master-{{ $index }}.${SEAWEEDFS_FULLNAME}-master:{{ $.Values.master.port }}{{ if lt $index (sub ($.Values.master.replicas | int) 1) }},{{ end }}{{ end }}
           volumeMounts:
-            - name: seaweedfs-volume-storage
-              mountPath: "/data/"
-            - name: seaweedfs-volume-log-volume
+            - name: data
+              mountPath: "{{ .Values.volume.dir }}/"
+            {{- if .Values.volume.dir_idx }}
+            - name: idx
+              mountPath: "{{ .Values.volume.dir_idx }}/"
+            {{- end }}
+            - name: logs
               mountPath: "/logs/"
             {{- if .Values.global.enableSecurity }}
             - name: security-config
@@ -173,15 +190,27 @@ spec:
           resources:
             {{ tpl .Values.volume.resources . | nindent 12 | trim }}
           {{- end }}
+      {{- $hostpath_exists := include "volume.hostpath_exists" . -}}
+      {{- if $hostpath_exists }}
       volumes:
-        - name: seaweedfs-volume-log-volume
-          hostPath:
-            path: /storage/logs/seaweedfs/volume
-            type: DirectoryOrCreate
-        - name: seaweedfs-volume-storage
+        {{- if eq .Values.volume.data.type "hostPath" }}
+        - name: data
           hostPath:
             path: /storage/object_store/
             type: DirectoryOrCreate
+        {{- end }}
+        {{- if and (eq .Values.volume.idx.type "hostPath") .Values.volume.dir_idx }}
+        - name: idx
+          hostPath:
+            path: /ssd/seaweedfs-volume-idx/
+            type: DirectoryOrCreate
+        {{- end }}
+        {{- if eq .Values.volume.logs.type "hostPath" }}
+        - name: logs
+          hostPath:
+            path: /storage/logs/seaweedfs/volume
+            type: DirectoryOrCreate
+        {{- end }}
         {{- if .Values.global.enableSecurity }}
         - name: security-config
           configMap:
@@ -205,8 +234,43 @@ spec:
       {{- if .Values.volume.extraVolumes }}
         {{ tpl .Values.volume.extraVolumes . | indent 8 | trim }}
       {{- end }}
+      {{- end }}
       {{- if .Values.volume.nodeSelector }}
       nodeSelector:
         {{ tpl .Values.volume.nodeSelector . | indent 8 | trim }}
       {{- end }}
+  {{- $pvc_exists := include "volume.pvc_exists" . -}}
+  {{- if $pvc_exists }}
+  volumeClaimTemplates:
+    {{- if eq .Values.volume.data.type "persistentVolumeClaim"}}
+    - metadata:
+        name: data
+      spec:
+        accessModes: [ "ReadWriteOnce" ]
+        storageClassName: {{ .Values.volume.data.storageClass }}
+        resources:
+          requests:
+            storage: {{ .Values.volume.data.size }}
+    {{- end }}
+    {{- if and (eq .Values.volume.idx.type "persistentVolumeClaim") .Values.volume.dir_idx }}
+    - metadata:
+        name: idx
+      spec:
+        accessModes: [ "ReadWriteOnce" ]
+        storageClassName: {{ .Values.volume.idx.storageClass }}
+        resources:
+          requests:
+            storage: {{ .Values.volume.idx.size }}
+    {{- end }}
+    {{- if eq .Values.volume.logs.type "persistentVolumeClaim" }}
+    - metadata:
+        name: logs
+      spec:
+        accessModes: [ "ReadWriteOnce" ]
+        storageClassName: {{ .Values.volume.logs.storageClass }}
+        resources:
+          requests:
+            storage: {{ .Values.volume.logs.size }}
+    {{- end }}
+      {{- end }}
 {{- end }}

k8s/seaweedfs/values.yaml

@@ -4,7 +4,7 @@ global:
   registry: ""
   repository: ""
   imageName: chrislusf/seaweedfs
-  # imageTag: "2.27" - started using {.Chart.appVersion}
+  # imageTag: "2.38" - started using {.Chart.appVersion}
   imagePullPolicy: IfNotPresent
   imagePullSecrets: imagepullsecret
   restartPolicy: Always
@@ -138,6 +138,24 @@ volume:
   # minimum free disk space(in percents). If free disk space lower this value - all volumes marks as ReadOnly
   minFreeSpacePercent: 7
 
+# can use ANY storage-class , example with local-path-provisner
+#  data:
+#    type: "persistentVolumeClaim"
+#    size: "24Ti"
+#    storageClass: "local-path-provisioner"
+  data:
+    type: "hostPath"
+    size: ""
+    storageClass: ""
+  idx:
+    type: "hostPath"
+    size: ""
+    storageClass: ""
+
+  logs:
+    type: "hostPath"
+    size: ""
+    storageClass: ""
 
   # limit background compaction or copying speed in mega bytes per second
   compactionMBps: "50"

other/java/client/pom.xml

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.6.1</version>
+    <version>1.6.4</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>
@@ -17,7 +17,7 @@
         <protobuf.version>3.9.1</protobuf.version>
         <!-- follow https://github.com/grpc/grpc-java -->
         <grpc.version>1.23.0</grpc.version>
-        <guava.version>28.0-jre</guava.version>
+        <guava.version>30.0-jre</guava.version>
     </properties>
 
     <dependencies>

other/java/client/pom.xml.deploy

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.6.1</version>
+    <version>1.6.4</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>

other/java/client/pom_debug.xml

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.6.1</version>
+    <version>1.6.4</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>

other/java/client/src/main/java/seaweedfs/client/ByteBufferPool.java

@@ -3,6 +3,7 @@ package seaweedfs.client;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import java.nio.Buffer;
 import java.nio.ByteBuffer;
 import java.util.ArrayList;
 import java.util.List;
@@ -34,7 +35,7 @@ public class ByteBufferPool {
     }
 
     public static synchronized void release(ByteBuffer obj) {
-        obj.clear();
+        ((Buffer)obj).clear();
         bufferList.add(0, obj);
     }
 

other/java/client/src/main/java/seaweedfs/client/FilerClient.java

@@ -94,7 +94,7 @@ public class FilerClient extends FilerGrpcClient {
             return true;
         }
         File pathFile = new File(path);
-        String parent = pathFile.getParent();
+        String parent = pathFile.getParent().replace('\\','/');
         String name = pathFile.getName();
 
         mkdirs(parent, mode, uid, gid, userName, groupNames);
@@ -115,11 +115,11 @@ public class FilerClient extends FilerGrpcClient {
     public boolean mv(String oldPath, String newPath) {
 
         File oldPathFile = new File(oldPath);
-        String oldParent = oldPathFile.getParent();
+        String oldParent = oldPathFile.getParent().replace('\\','/');
         String oldName = oldPathFile.getName();
 
         File newPathFile = new File(newPath);
-        String newParent = newPathFile.getParent();
+        String newParent = newPathFile.getParent().replace('\\','/');
         String newName = newPathFile.getName();
 
         return atomicRenameEntry(oldParent, oldName, newParent, newName);
@@ -129,7 +129,7 @@ public class FilerClient extends FilerGrpcClient {
     public boolean rm(String path, boolean isRecursive, boolean ignoreRecusiveError) {
 
         File pathFile = new File(path);
-        String parent = pathFile.getParent();
+        String parent = pathFile.getParent().replace('\\','/');
         String name = pathFile.getName();
 
         return deleteEntry(
@@ -148,7 +148,7 @@ public class FilerClient extends FilerGrpcClient {
     public boolean touch(String path, int mode, int uid, int gid, String userName, String[] groupNames) {
 
         File pathFile = new File(path);
-        String parent = pathFile.getParent();
+        String parent = pathFile.getParent().replace('\\','/');
         String name = pathFile.getName();
 
         FilerProto.Entry entry = lookupEntry(parent, name);

other/java/client/src/main/java/seaweedfs/client/SeaweedOutputStream.java

@@ -8,6 +8,7 @@ import org.slf4j.LoggerFactory;
 import java.io.IOException;
 import java.io.InterruptedIOException;
 import java.io.OutputStream;
+import java.nio.Buffer;
 import java.nio.ByteBuffer;
 import java.util.concurrent.*;
 
@@ -217,7 +218,7 @@ public class SeaweedOutputStream extends OutputStream {
 
     private synchronized int submitWriteBufferToService(final ByteBuffer bufferToWrite, final long writePosition) throws IOException {
 
-        bufferToWrite.flip();
+        ((Buffer)bufferToWrite).flip();
         int bytesLength = bufferToWrite.limit() - bufferToWrite.position();
 
         if (threadExecutor.getQueue().size() >= maxConcurrentRequestCount) {

other/java/examples/pom.xml

@@ -11,13 +11,13 @@
         <dependency>
             <groupId>com.github.chrislusf</groupId>
             <artifactId>seaweedfs-client</artifactId>
-            <version>1.6.1</version>
+            <version>1.6.4</version>
             <scope>compile</scope>
         </dependency>
         <dependency>
             <groupId>com.github.chrislusf</groupId>
             <artifactId>seaweedfs-hadoop2-client</artifactId>
-            <version>1.6.1</version>
+            <version>1.6.4</version>
             <scope>compile</scope>
         </dependency>
         <dependency>

other/java/hdfs2/dependency-reduced-pom.xml

@@ -301,7 +301,7 @@
     </snapshotRepository>
   </distributionManagement>
   <properties>
-    <seaweedfs.client.version>1.6.1</seaweedfs.client.version>
+    <seaweedfs.client.version>1.6.4</seaweedfs.client.version>
     <hadoop.version>2.9.2</hadoop.version>
   </properties>
 </project>

other/java/hdfs2/pom.xml

@@ -5,7 +5,7 @@
     <modelVersion>4.0.0</modelVersion>
 
     <properties>
-        <seaweedfs.client.version>1.6.1</seaweedfs.client.version>
+        <seaweedfs.client.version>1.6.4</seaweedfs.client.version>
         <hadoop.version>2.9.2</hadoop.version>
     </properties>
 

other/java/hdfs3/dependency-reduced-pom.xml

@@ -309,7 +309,7 @@
     </snapshotRepository>
   </distributionManagement>
   <properties>
-    <seaweedfs.client.version>1.6.1</seaweedfs.client.version>
+    <seaweedfs.client.version>1.6.4</seaweedfs.client.version>
     <hadoop.version>3.1.1</hadoop.version>
   </properties>
 </project>

other/java/hdfs3/pom.xml

@@ -5,7 +5,7 @@
     <modelVersion>4.0.0</modelVersion>
 
     <properties>
-        <seaweedfs.client.version>1.6.1</seaweedfs.client.version>
+        <seaweedfs.client.version>1.6.4</seaweedfs.client.version>
         <hadoop.version>3.1.1</hadoop.version>
     </properties>
 

test/random_access/pom.xml

@@ -8,7 +8,7 @@
     <version>1.0-SNAPSHOT</version>
 
     <properties>
-        <guava.version>28.0-jre</guava.version>
+        <guava.version>30.0-jre</guava.version>
     </properties>
 
     <dependencies>

weed/Makefile

@@ -33,3 +33,7 @@ debug_webdav:
 debug_s3:
 	go build -gcflags="all=-N -l"
 	dlv --listen=:2345 --headless=true --api-version=2 --accept-multiclient exec weed -- -v=4 s3
+
+debug_filer_copy:
+	go build -gcflags="all=-N -l"
+	dlv --listen=:2345 --headless=true --api-version=2 --accept-multiclient exec weed -- -v=4 filer.backup -filer=localhost:8888 -filerProxy -timeAgo=10h

weed/command/benchmark.go

@@ -41,6 +41,7 @@ type BenchmarkOptions struct {
 	grpcDialOption   grpc.DialOption
 	masterClient     *wdclient.MasterClient
 	fsync            *bool
+	useTcp           *bool
 }
 
 var (
@@ -63,10 +64,11 @@ func init() {
 	b.sequentialRead = cmdBenchmark.Flag.Bool("readSequentially", false, "randomly read by ids from \"-list\" specified file")
 	b.collection = cmdBenchmark.Flag.String("collection", "benchmark", "write data to this collection")
 	b.replication = cmdBenchmark.Flag.String("replication", "000", "replication type")
-	b.diskType = cmdBenchmark.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
+	b.diskType = cmdBenchmark.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	b.cpuprofile = cmdBenchmark.Flag.String("cpuprofile", "", "cpu profile output file")
 	b.maxCpu = cmdBenchmark.Flag.Int("maxCpu", 0, "maximum number of CPUs. 0 means all available CPUs")
 	b.fsync = cmdBenchmark.Flag.Bool("fsync", false, "flush data to disk after write")
+	b.useTcp = cmdBenchmark.Flag.Bool("useTcp", false, "send data via tcp")
 	sharedBytes = make([]byte, 1024)
 }
 
@@ -223,6 +225,8 @@ func writeFiles(idChan chan int, fileIdLineChan chan string, s *stat) {
 
 	random := rand.New(rand.NewSource(time.Now().UnixNano()))
 
+	volumeTcpClient := wdclient.NewVolumeTcpClient()
+
 	for id := range idChan {
 		start := time.Now()
 		fileSize := int64(*b.fileSize + random.Intn(64))
@@ -243,7 +247,15 @@ func writeFiles(idChan chan int, fileIdLineChan chan string, s *stat) {
 			if !isSecure && assignResult.Auth != "" {
 				isSecure = true
 			}
-			if _, err := fp.Upload(0, b.masterClient.GetMaster, false, assignResult.Auth, b.grpcDialOption); err == nil {
+			if *b.useTcp {
+				if uploadByTcp(volumeTcpClient, fp) {
+					fileIdLineChan <- fp.Fid
+					s.completed++
+					s.transferred += fileSize
+				} else {
+					s.failed++
+				}
+			} else if _, err := fp.Upload(0, b.masterClient.GetMaster, false, assignResult.Auth, b.grpcDialOption); err == nil {
 				if random.Intn(100) < *b.deletePercentage {
 					s.total++
 					delayedDeleteChan <- &delayedFile{time.Now().Add(time.Second), fp}
@@ -293,7 +305,7 @@ func readFiles(fileIdLineChan chan string, s *stat) {
 		}
 		var bytes []byte
 		for _, url := range urls {
-			bytes, _, err = util.FastGet(url)
+			bytes, _, err = util.Get(url)
 			if err == nil {
 				break
 			}
@@ -329,6 +341,17 @@ func writeFileIds(fileName string, fileIdLineChan chan string, finishChan chan b
 	}
 }
 
+func uploadByTcp(volumeTcpClient *wdclient.VolumeTcpClient, fp *operation.FilePart) bool {
+
+	err := volumeTcpClient.PutFileChunk(fp.Server, fp.Fid, uint32(fp.FileSize), fp.Reader)
+	if err != nil {
+		glog.Errorf("upload chunk err: %v", err)
+		return false
+	}
+
+	return true
+}
+
 func readFileIds(fileName string, fileIdLineChan chan string) {
 	file, err := os.Open(fileName) // For read access.
 	if err != nil {

weed/command/command.go

@@ -15,7 +15,9 @@ var Commands = []*Command{
 	cmdDownload,
 	cmdExport,
 	cmdFiler,
+	cmdFilerBackup,
 	cmdFilerCat,
+	cmdFilerMetaBackup,
 	cmdFilerMetaTail,
 	cmdFilerReplicate,
 	cmdFilerSynchronize,

weed/command/filer.go

@@ -47,6 +47,7 @@ type FilerOptions struct {
 	metricsHttpPort         *int
 	saveToFilerLimit        *int
 	defaultLevelDbDirectory *string
+	concurrentUploadLimitMB *int
 }
 
 func init() {
@@ -54,12 +55,12 @@ func init() {
 	f.masters = cmdFiler.Flag.String("master", "localhost:9333", "comma-separated master servers")
 	f.collection = cmdFiler.Flag.String("collection", "", "all data will be stored in this default collection")
 	f.ip = cmdFiler.Flag.String("ip", util.DetectedHostAddress(), "filer server http listen ip address")
-	f.bindIp = cmdFiler.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to")
+	f.bindIp = cmdFiler.Flag.String("ip.bind", "", "ip address to bind to")
 	f.port = cmdFiler.Flag.Int("port", 8888, "filer server http listen port")
 	f.publicPort = cmdFiler.Flag.Int("port.readonly", 0, "readonly port opened to public")
 	f.defaultReplicaPlacement = cmdFiler.Flag.String("defaultReplicaPlacement", "", "default replication type. If not specified, use master setting.")
 	f.disableDirListing = cmdFiler.Flag.Bool("disableDirListing", false, "turn off directory listing")
-	f.maxMB = cmdFiler.Flag.Int("maxMB", 32, "split files larger than the limit")
+	f.maxMB = cmdFiler.Flag.Int("maxMB", 4, "split files larger than the limit")
 	f.dirListingLimit = cmdFiler.Flag.Int("dirListLimit", 100000, "limit sub dir listing size")
 	f.dataCenter = cmdFiler.Flag.String("dataCenter", "", "prefer to read and write to volumes in this data center")
 	f.rack = cmdFiler.Flag.String("rack", "", "prefer to write to volumes in this rack")
@@ -69,6 +70,7 @@ func init() {
 	f.metricsHttpPort = cmdFiler.Flag.Int("metricsPort", 0, "Prometheus metrics listen port")
 	f.saveToFilerLimit = cmdFiler.Flag.Int("saveToFilerLimit", 0, "files smaller than this limit will be saved in filer store")
 	f.defaultLevelDbDirectory = cmdFiler.Flag.String("defaultStoreDir", ".", "if filer.toml is empty, use an embedded filer store in the directory")
+	f.concurrentUploadLimitMB = cmdFiler.Flag.Int("concurrentUploadLimitMB", 128, "limit total concurrent upload size")
 
 	// start s3 on filer
 	filerStartS3 = cmdFiler.Flag.Bool("s3", false, "whether to start S3 gateway")
@@ -84,7 +86,7 @@ func init() {
 	filerWebDavOptions.port = cmdFiler.Flag.Int("webdav.port", 7333, "webdav server http listen port")
 	filerWebDavOptions.collection = cmdFiler.Flag.String("webdav.collection", "", "collection to create the files")
 	filerWebDavOptions.replication = cmdFiler.Flag.String("webdav.replication", "", "replication to create the files")
-	filerWebDavOptions.disk = cmdFiler.Flag.String("webdav.disk", "", "[hdd|ssd] hard drive or solid state drive")
+	filerWebDavOptions.disk = cmdFiler.Flag.String("webdav.disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	filerWebDavOptions.tlsPrivateKey = cmdFiler.Flag.String("webdav.key.file", "", "path to the TLS private key file")
 	filerWebDavOptions.tlsCertificate = cmdFiler.Flag.String("webdav.cert.file", "", "path to the TLS certificate file")
 	filerWebDavOptions.cacheDir = cmdFiler.Flag.String("webdav.cacheDir", os.TempDir(), "local cache directory for file chunks")
@@ -159,21 +161,22 @@ func (fo *FilerOptions) startFiler() {
 	}
 
 	fs, nfs_err := weed_server.NewFilerServer(defaultMux, publicVolumeMux, &weed_server.FilerOption{
-		Masters:            strings.Split(*fo.masters, ","),
-		Collection:         *fo.collection,
-		DefaultReplication: *fo.defaultReplicaPlacement,
-		DisableDirListing:  *fo.disableDirListing,
-		MaxMB:              *fo.maxMB,
-		DirListingLimit:    *fo.dirListingLimit,
-		DataCenter:         *fo.dataCenter,
-		Rack:               *fo.rack,
-		DefaultLevelDbDir:  defaultLevelDbDirectory,
-		DisableHttp:        *fo.disableHttp,
-		Host:               *fo.ip,
-		Port:               uint32(*fo.port),
-		Cipher:             *fo.cipher,
-		SaveToFilerLimit:   *fo.saveToFilerLimit,
-		Filers:             peers,
+		Masters:               strings.Split(*fo.masters, ","),
+		Collection:            *fo.collection,
+		DefaultReplication:    *fo.defaultReplicaPlacement,
+		DisableDirListing:     *fo.disableDirListing,
+		MaxMB:                 *fo.maxMB,
+		DirListingLimit:       *fo.dirListingLimit,
+		DataCenter:            *fo.dataCenter,
+		Rack:                  *fo.rack,
+		DefaultLevelDbDir:     defaultLevelDbDirectory,
+		DisableHttp:           *fo.disableHttp,
+		Host:                  *fo.ip,
+		Port:                  uint32(*fo.port),
+		Cipher:                *fo.cipher,
+		SaveToFilerLimit:      int64(*fo.saveToFilerLimit),
+		Filers:                peers,
+		ConcurrentUploadLimit: int64(*fo.concurrentUploadLimitMB) * 1024 * 1024,
 	})
 	if nfs_err != nil {
 		glog.Fatalf("Filer startup error: %v", nfs_err)

weed/command/filer_backup.go

@@ -0,0 +1,157 @@
+package command
+
+import (
+	"context"
+	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/pb"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/replication/source"
+	"github.com/chrislusf/seaweedfs/weed/security"
+	"github.com/chrislusf/seaweedfs/weed/util"
+	"google.golang.org/grpc"
+	"io"
+	"time"
+)
+
+type FilerBackupOptions struct {
+	isActivePassive *bool
+	filer           *string
+	path            *string
+	debug           *bool
+	proxyByFiler    *bool
+	timeAgo         *time.Duration
+}
+
+var (
+	filerBackupOptions FilerBackupOptions
+)
+
+func init() {
+	cmdFilerBackup.Run = runFilerBackup // break init cycle
+	filerBackupOptions.filer = cmdFilerBackup.Flag.String("filer", "localhost:8888", "filer of one SeaweedFS cluster")
+	filerBackupOptions.path = cmdFilerBackup.Flag.String("filerPath", "/", "directory to sync on filer")
+	filerBackupOptions.proxyByFiler = cmdFilerBackup.Flag.Bool("filerProxy", false, "read and write file chunks by filer instead of volume servers")
+	filerBackupOptions.debug = cmdFilerBackup.Flag.Bool("debug", false, "debug mode to print out received files")
+	filerBackupOptions.timeAgo = cmdFilerBackup.Flag.Duration("timeAgo", 0, "start time before now. \"300ms\", \"1.5h\" or \"2h45m\". Valid time units are \"ns\", \"us\" (or \"µs\"), \"ms\", \"s\", \"m\", \"h\"")
+}
+
+var cmdFilerBackup = &Command{
+	UsageLine: "filer.backup -filer=<filerHost>:<filerPort> ",
+	Short:     "resume-able continuously replicate files from a SeaweedFS cluster to another location defined in replication.toml",
+	Long: `resume-able continuously replicate files from a SeaweedFS cluster to another location defined in replication.toml
+
+	filer.backup listens on filer notifications. If any file is updated, it will fetch the updated content,
+	and write to the destination. This is to replace filer.replicate command since additional message queue is not needed.
+
+	If restarted and "-timeAgo" is not set, the synchronization will resume from the previous checkpoints, persisted every minute.
+	A fresh sync will start from the earliest metadata logs. To reset the checkpoints, just set "-timeAgo" to a high value.
+
+`,
+}
+
+func runFilerBackup(cmd *Command, args []string) bool {
+
+	grpcDialOption := security.LoadClientTLS(util.GetViper(), "grpc.client")
+
+	util.LoadConfiguration("security", false)
+	util.LoadConfiguration("replication", true)
+
+	for {
+		err := doFilerBackup(grpcDialOption, &filerBackupOptions)
+		if err != nil {
+			glog.Errorf("backup from %s: %v", *filerBackupOptions.filer, err)
+			time.Sleep(1747 * time.Millisecond)
+		}
+	}
+
+	return true
+}
+
+const (
+	BackupKeyPrefix = "backup."
+)
+
+func doFilerBackup(grpcDialOption grpc.DialOption, backupOption *FilerBackupOptions) error {
+
+	// find data sink
+	config := util.GetViper()
+	dataSink := findSink(config)
+	if dataSink == nil {
+		return fmt.Errorf("no data sink configured in replication.toml")
+	}
+
+	sourceFiler := *backupOption.filer
+	sourcePath := *backupOption.path
+	timeAgo := *backupOption.timeAgo
+	targetPath := dataSink.GetSinkToDirectory()
+	debug := *backupOption.debug
+
+	// get start time for the data sink
+	startFrom := time.Unix(0, 0)
+	sinkId := util.HashStringToLong(dataSink.GetName() + dataSink.GetSinkToDirectory())
+	if timeAgo.Milliseconds() == 0 {
+		lastOffsetTsNs, err := getOffset(grpcDialOption, sourceFiler, BackupKeyPrefix, int32(sinkId))
+		if err != nil {
+			glog.V(0).Infof("starting from %v", startFrom)
+		} else {
+			startFrom = time.Unix(0, lastOffsetTsNs)
+			glog.V(0).Infof("resuming from %v", startFrom)
+		}
+	} else {
+		startFrom = time.Now().Add(-timeAgo)
+		glog.V(0).Infof("start time is set to %v", startFrom)
+	}
+
+	// create filer sink
+	filerSource := &source.FilerSource{}
+	filerSource.DoInitialize(sourceFiler, pb.ServerToGrpcAddress(sourceFiler), sourcePath, *backupOption.proxyByFiler)
+	dataSink.SetSourceFiler(filerSource)
+
+	processEventFn := genProcessFunction(sourcePath, targetPath, dataSink, debug)
+
+	return pb.WithFilerClient(sourceFiler, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
+
+		ctx, cancel := context.WithCancel(context.Background())
+		defer cancel()
+
+		stream, err := client.SubscribeMetadata(ctx, &filer_pb.SubscribeMetadataRequest{
+			ClientName: "backup_" + dataSink.GetName(),
+			PathPrefix: sourcePath,
+			SinceNs:    startFrom.UnixNano(),
+		})
+		if err != nil {
+			return fmt.Errorf("listen: %v", err)
+		}
+
+		var counter int64
+		var lastWriteTime time.Time
+		for {
+			resp, listenErr := stream.Recv()
+
+			if listenErr == io.EOF {
+				return nil
+			}
+			if listenErr != nil {
+				return listenErr
+			}
+
+			if err := processEventFn(resp); err != nil {
+				return fmt.Errorf("processEventFn: %v", err)
+			}
+
+			counter++
+			if lastWriteTime.Add(3 * time.Second).Before(time.Now()) {
+				glog.V(0).Infof("backup %s progressed to %v %0.2f/sec", sourceFiler, time.Unix(0, resp.TsNs), float64(counter)/float64(3))
+				counter = 0
+				lastWriteTime = time.Now()
+				if err := setOffset(grpcDialOption, sourceFiler, BackupKeyPrefix, int32(sinkId), resp.TsNs); err != nil {
+					return fmt.Errorf("setOffset: %v", err)
+				}
+			}
+
+		}
+
+	})
+
+}

weed/command/filer_cat.go

@@ -110,7 +110,7 @@ func runFilerCat(cmd *Command, args []string) bool {
 
 		filerCat.filerClient = client
 
-		return filer.StreamContent(&filerCat, writer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64)
+		return filer.StreamContent(&filerCat, writer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64, false)
 
 	})
 

weed/command/filer_copy.go

@@ -55,8 +55,8 @@ func init() {
 	copy.replication = cmdCopy.Flag.String("replication", "", "replication type")
 	copy.collection = cmdCopy.Flag.String("collection", "", "optional collection name")
 	copy.ttl = cmdCopy.Flag.String("ttl", "", "time to live, e.g.: 1m, 1h, 1d, 1M, 1y")
-	copy.diskType = cmdCopy.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
-	copy.maxMB = cmdCopy.Flag.Int("maxMB", 32, "split files larger than the limit")
+	copy.diskType = cmdCopy.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
+	copy.maxMB = cmdCopy.Flag.Int("maxMB", 4, "split files larger than the limit")
 	copy.concurrenctFiles = cmdCopy.Flag.Int("c", 8, "concurrent file copy goroutines")
 	copy.concurrenctChunks = cmdCopy.Flag.Int("concurrentChunks", 8, "concurrent chunk copy goroutines for each file")
 }

weed/command/filer_meta_backup.go

@@ -0,0 +1,268 @@
+package command
+
+import (
+	"context"
+	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/filer"
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/spf13/viper"
+	"google.golang.org/grpc"
+	"io"
+	"reflect"
+	"time"
+
+	"github.com/chrislusf/seaweedfs/weed/pb"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/security"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+var (
+	metaBackup FilerMetaBackupOptions
+)
+
+type FilerMetaBackupOptions struct {
+	grpcDialOption    grpc.DialOption
+	filerAddress      *string
+	filerDirectory    *string
+	restart           *bool
+	backupFilerConfig *string
+
+	store filer.FilerStore
+}
+
+func init() {
+	cmdFilerMetaBackup.Run = runFilerMetaBackup // break init cycle
+	metaBackup.filerAddress = cmdFilerMetaBackup.Flag.String("filer", "localhost:8888", "filer hostname:port")
+	metaBackup.filerDirectory = cmdFilerMetaBackup.Flag.String("filerDir", "/", "a folder on the filer")
+	metaBackup.restart = cmdFilerMetaBackup.Flag.Bool("restart", false, "copy the full metadata before async incremental backup")
+	metaBackup.backupFilerConfig = cmdFilerMetaBackup.Flag.String("config", "", "path to filer.toml specifying backup filer store")
+}
+
+var cmdFilerMetaBackup = &Command{
+	UsageLine: "filer.meta.backup [-filer=localhost:8888] [-filerDir=/] [-restart] -config=/path/to/backup_filer.toml",
+	Short:     "continuously backup filer meta data changes to anther filer store specified in a backup_filer.toml",
+	Long: `continuously backup filer meta data changes. 
+The backup writes to another filer store specified in a backup_filer.toml.
+
+	weed filer.meta.backup -config=/path/to/backup_filer.toml -filer="localhost:8888"
+	weed filer.meta.backup -config=/path/to/backup_filer.toml -filer="localhost:8888" -restart
+
+  `,
+}
+
+func runFilerMetaBackup(cmd *Command, args []string) bool {
+
+	metaBackup.grpcDialOption = security.LoadClientTLS(util.GetViper(), "grpc.client")
+
+	// load backup_filer.toml
+	v := viper.New()
+	v.SetConfigFile(*metaBackup.backupFilerConfig)
+
+	if err := v.ReadInConfig(); err != nil { // Handle errors reading the config file
+		glog.Fatalf("Failed to load %s file.\nPlease use this command to generate the a %s.toml file\n"+
+			"    weed scaffold -config=%s -output=.\n\n\n",
+			*metaBackup.backupFilerConfig, "backup_filer", "filer")
+	}
+
+	if err := metaBackup.initStore(v); err != nil {
+		glog.V(0).Infof("init backup filer store: %v", err)
+		return true
+	}
+
+	missingPreviousBackup := false
+	_, err := metaBackup.getOffset()
+	if err != nil {
+		missingPreviousBackup = true
+	}
+
+	if *metaBackup.restart || missingPreviousBackup {
+		glog.V(0).Infof("traversing metadata tree...")
+		startTime := time.Now()
+		if err := metaBackup.traverseMetadata(); err != nil {
+			glog.Errorf("traverse meta data: %v", err)
+			return true
+		}
+		glog.V(0).Infof("metadata copied up to %v", startTime)
+		if err := metaBackup.setOffset(startTime); err != nil {
+			startTime = time.Now()
+		}
+	}
+
+	for {
+		err := metaBackup.streamMetadataBackup()
+		if err != nil {
+			glog.Errorf("filer meta backup from %s: %v", *metaBackup.filerAddress, err)
+			time.Sleep(1747 * time.Millisecond)
+		}
+	}
+
+	return true
+}
+
+func (metaBackup *FilerMetaBackupOptions) initStore(v *viper.Viper) error {
+	// load configuration for default filer store
+	hasDefaultStoreConfigured := false
+	for _, store := range filer.Stores {
+		if v.GetBool(store.GetName() + ".enabled") {
+			store = reflect.New(reflect.ValueOf(store).Elem().Type()).Interface().(filer.FilerStore)
+			if err := store.Initialize(v, store.GetName()+"."); err != nil {
+				glog.Fatalf("failed to initialize store for %s: %+v", store.GetName(), err)
+			}
+			glog.V(0).Infof("configured filer store to %s", store.GetName())
+			hasDefaultStoreConfigured = true
+			metaBackup.store = filer.NewFilerStoreWrapper(store)
+			break
+		}
+	}
+	if !hasDefaultStoreConfigured {
+		return fmt.Errorf("no filer store enabled in %s", v.ConfigFileUsed())
+	}
+
+	return nil
+}
+
+func (metaBackup *FilerMetaBackupOptions) traverseMetadata() (err error) {
+	var saveErr error
+
+	traverseErr := filer_pb.TraverseBfs(metaBackup, util.FullPath(*metaBackup.filerDirectory), func(parentPath util.FullPath, entry *filer_pb.Entry) {
+
+		println("+", parentPath.Child(entry.Name))
+		if err := metaBackup.store.InsertEntry(context.Background(), filer.FromPbEntry(string(parentPath), entry)); err != nil {
+			saveErr = fmt.Errorf("insert entry error: %v\n", err)
+			return
+		}
+
+	})
+
+	if traverseErr != nil {
+		return fmt.Errorf("traverse: %v", traverseErr)
+	}
+	return saveErr
+}
+
+var (
+	MetaBackupKey = []byte("metaBackup")
+)
+
+func (metaBackup *FilerMetaBackupOptions) streamMetadataBackup() error {
+
+	startTime, err := metaBackup.getOffset()
+	if err != nil {
+		startTime = time.Now()
+	}
+	glog.V(0).Infof("streaming from %v", startTime)
+
+	store := metaBackup.store
+
+	eachEntryFunc := func(resp *filer_pb.SubscribeMetadataResponse) error {
+
+		ctx := context.Background()
+		message := resp.EventNotification
+
+		if message.OldEntry == nil && message.NewEntry == nil {
+			return nil
+		}
+		if message.OldEntry == nil && message.NewEntry != nil {
+			println("+", util.FullPath(message.NewParentPath).Child(message.NewEntry.Name))
+			entry := filer.FromPbEntry(message.NewParentPath, message.NewEntry)
+			return store.InsertEntry(ctx, entry)
+		}
+		if message.OldEntry != nil && message.NewEntry == nil {
+			println("-", util.FullPath(resp.Directory).Child(message.OldEntry.Name))
+			return store.DeleteEntry(ctx, util.FullPath(resp.Directory).Child(message.OldEntry.Name))
+		}
+		if message.OldEntry != nil && message.NewEntry != nil {
+			if resp.Directory == message.NewParentPath && message.OldEntry.Name == message.NewEntry.Name {
+				println("~", util.FullPath(message.NewParentPath).Child(message.NewEntry.Name))
+				entry := filer.FromPbEntry(message.NewParentPath, message.NewEntry)
+				return store.UpdateEntry(ctx, entry)
+			}
+			println("-", util.FullPath(resp.Directory).Child(message.OldEntry.Name))
+			if err := store.DeleteEntry(ctx, util.FullPath(resp.Directory).Child(message.OldEntry.Name)); err != nil {
+				return err
+			}
+			println("+", util.FullPath(message.NewParentPath).Child(message.NewEntry.Name))
+			return store.InsertEntry(ctx, filer.FromPbEntry(message.NewParentPath, message.NewEntry))
+		}
+
+		return nil
+	}
+
+	tailErr := pb.WithFilerClient(*metaBackup.filerAddress, metaBackup.grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
+
+		ctx, cancel := context.WithCancel(context.Background())
+		defer cancel()
+
+		stream, err := client.SubscribeMetadata(ctx, &filer_pb.SubscribeMetadataRequest{
+			ClientName: "meta_backup",
+			PathPrefix: *metaBackup.filerDirectory,
+			SinceNs:    startTime.UnixNano(),
+		})
+		if err != nil {
+			return fmt.Errorf("listen: %v", err)
+		}
+
+		var counter int64
+		var lastWriteTime time.Time
+		for {
+			resp, listenErr := stream.Recv()
+			if listenErr == io.EOF {
+				return nil
+			}
+			if listenErr != nil {
+				return listenErr
+			}
+			if err = eachEntryFunc(resp); err != nil {
+				return err
+			}
+
+			counter++
+			if lastWriteTime.Add(3 * time.Second).Before(time.Now()) {
+				glog.V(0).Infof("meta backup %s progressed to %v %0.2f/sec", *metaBackup.filerAddress, time.Unix(0, resp.TsNs), float64(counter)/float64(3))
+				counter = 0
+				lastWriteTime = time.Now()
+				if err2 := metaBackup.setOffset(lastWriteTime); err2 != nil {
+					return err2
+				}
+			}
+
+		}
+
+	})
+	return tailErr
+}
+
+func (metaBackup *FilerMetaBackupOptions) getOffset() (lastWriteTime time.Time, err error) {
+	value, err := metaBackup.store.KvGet(context.Background(), MetaBackupKey)
+	if err != nil {
+		return
+	}
+	tsNs := util.BytesToUint64(value)
+
+	return time.Unix(0, int64(tsNs)), nil
+}
+
+func (metaBackup *FilerMetaBackupOptions) setOffset(lastWriteTime time.Time) error {
+	valueBuf := make([]byte, 8)
+	util.Uint64toBytes(valueBuf, uint64(lastWriteTime.UnixNano()))
+
+	if err := metaBackup.store.KvPut(context.Background(), MetaBackupKey, valueBuf); err != nil {
+		return err
+	}
+	return nil
+}
+
+var _ = filer_pb.FilerClient(&FilerMetaBackupOptions{})
+
+func (metaBackup *FilerMetaBackupOptions) WithFilerClient(fn func(filer_pb.SeaweedFilerClient) error) error {
+
+	return pb.WithFilerClient(*metaBackup.filerAddress, metaBackup.grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
+		return fn(client)
+	})
+
+}
+
+func (metaBackup *FilerMetaBackupOptions) AdjustedUrl(location *filer_pb.Location) string {
+	return location.Url
+}

weed/command/filer_meta_tail.go

@@ -23,9 +23,9 @@ func init() {
 }
 
 var cmdFilerMetaTail = &Command{
-	UsageLine: "filer.meta.tail [-filer=localhost:8888] [-target=/]",
-	Short:     "see recent changes on a filer",
-	Long: `See recent changes on a filer.
+	UsageLine: "filer.meta.tail [-filer=localhost:8888] [-pathPrefix=/]",
+	Short:     "see continuous changes on a filer",
+	Long: `See continuous changes on a filer.
 
 	weed filer.meta.tail -timeAgo=30h | grep truncate
 	weed filer.meta.tail -timeAgo=30h | jq .
@@ -36,7 +36,7 @@ var cmdFilerMetaTail = &Command{
 
 var (
 	tailFiler   = cmdFilerMetaTail.Flag.String("filer", "localhost:8888", "filer hostname:port")
-	tailTarget  = cmdFilerMetaTail.Flag.String("pathPrefix", "/", "path to a folder or file, or common prefix for the folders or files on filer")
+	tailTarget  = cmdFilerMetaTail.Flag.String("pathPrefix", "/", "path to a folder or common prefix for the folders or files on filer")
 	tailStart   = cmdFilerMetaTail.Flag.Duration("timeAgo", 0, "start time before now. \"300ms\", \"1.5h\" or \"2h45m\". Valid time units are \"ns\", \"us\" (or \"µs\"), \"ms\", \"s\", \"m\", \"h\"")
 	tailPattern = cmdFilerMetaTail.Flag.String("pattern", "", "full path or just filename pattern, ex: \"/home/?opher\", \"*.pdf\", see https://golang.org/pkg/path/filepath/#Match ")
 	esServers   = cmdFilerMetaTail.Flag.String("es", "", "comma-separated elastic servers http://<host:port>")

weed/command/filer_replication.go

@@ -74,18 +74,7 @@ func runFilerReplicate(cmd *Command, args []string) bool {
 		}
 	}
 
-	var dataSink sink.ReplicationSink
-	for _, sk := range sink.Sinks {
-		if config.GetBool("sink." + sk.GetName() + ".enabled") {
-			if err := sk.Initialize(config, "sink."+sk.GetName()+"."); err != nil {
-				glog.Fatalf("Failed to initialize sink for %s: %+v",
-					sk.GetName(), err)
-			}
-			glog.V(0).Infof("Configure sink to %s", sk.GetName())
-			dataSink = sk
-			break
-		}
-	}
+	dataSink := findSink(config)
 
 	if dataSink == nil {
 		println("no data sink configured in replication.toml:")
@@ -135,6 +124,22 @@ func runFilerReplicate(cmd *Command, args []string) bool {
 
 }
 
+func findSink(config *util.ViperProxy) sink.ReplicationSink {
+	var dataSink sink.ReplicationSink
+	for _, sk := range sink.Sinks {
+		if config.GetBool("sink." + sk.GetName() + ".enabled") {
+			if err := sk.Initialize(config, "sink."+sk.GetName()+"."); err != nil {
+				glog.Fatalf("Failed to initialize sink for %s: %+v",
+					sk.GetName(), err)
+			}
+			glog.V(0).Infof("Configure sink to %s", sk.GetName())
+			dataSink = sk
+			break
+		}
+	}
+	return dataSink
+}
+
 func validateOneEnabledInput(config *util.ViperProxy) {
 	enabledInput := ""
 	for _, input := range sub.NotificationInputs {

weed/command/filer_sync.go

@@ -8,6 +8,7 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/pb"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
 	"github.com/chrislusf/seaweedfs/weed/replication"
+	"github.com/chrislusf/seaweedfs/weed/replication/sink"
 	"github.com/chrislusf/seaweedfs/weed/replication/sink/filersink"
 	"github.com/chrislusf/seaweedfs/weed/replication/source"
 	"github.com/chrislusf/seaweedfs/weed/security"
@@ -58,8 +59,8 @@ func init() {
 	syncOptions.bCollection = cmdFilerSynchronize.Flag.String("b.collection", "", "collection on filer B")
 	syncOptions.aTtlSec = cmdFilerSynchronize.Flag.Int("a.ttlSec", 0, "ttl in seconds on filer A")
 	syncOptions.bTtlSec = cmdFilerSynchronize.Flag.Int("b.ttlSec", 0, "ttl in seconds on filer B")
-	syncOptions.aDiskType = cmdFilerSynchronize.Flag.String("a.disk", "", "[hdd|ssd] hard drive or solid state drive on filer A")
-	syncOptions.bDiskType = cmdFilerSynchronize.Flag.String("b.disk", "", "[hdd|ssd] hard drive or solid state drive on filer B")
+	syncOptions.aDiskType = cmdFilerSynchronize.Flag.String("a.disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag on filer A")
+	syncOptions.bDiskType = cmdFilerSynchronize.Flag.String("b.disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag on filer B")
 	syncOptions.aProxyByFiler = cmdFilerSynchronize.Flag.Bool("a.filerProxy", false, "read and write file chunks by filer A instead of volume servers")
 	syncOptions.bProxyByFiler = cmdFilerSynchronize.Flag.Bool("b.filerProxy", false, "read and write file chunks by filer B instead of volume servers")
 	syncOptions.aDebug = cmdFilerSynchronize.Flag.Bool("a.debug", false, "debug mode to print out filer A received files")
@@ -137,7 +138,7 @@ func doSubscribeFilerMetaChanges(grpcDialOption grpc.DialOption, sourceFiler, so
 
 	// if first time, start from now
 	// if has previously synced, resume from that point of time
-	sourceFilerOffsetTsNs, err := readSyncOffset(grpcDialOption, targetFiler, sourceFilerSignature)
+	sourceFilerOffsetTsNs, err := getOffset(grpcDialOption, targetFiler, SyncKeyPrefix, sourceFilerSignature)
 	if err != nil {
 		return err
 	}
@@ -151,93 +152,17 @@ func doSubscribeFilerMetaChanges(grpcDialOption grpc.DialOption, sourceFiler, so
 	filerSink.DoInitialize(targetFiler, pb.ServerToGrpcAddress(targetFiler), targetPath, replicationStr, collection, ttlSec, diskType, grpcDialOption, sinkWriteChunkByFiler)
 	filerSink.SetSourceFiler(filerSource)
 
+	persistEventFn := genProcessFunction(sourcePath, targetPath, filerSink, debug)
+
 	processEventFn := func(resp *filer_pb.SubscribeMetadataResponse) error {
 		message := resp.EventNotification
-
-		var sourceOldKey, sourceNewKey util.FullPath
-		if message.OldEntry != nil {
-			sourceOldKey = util.FullPath(resp.Directory).Child(message.OldEntry.Name)
-		}
-		if message.NewEntry != nil {
-			sourceNewKey = util.FullPath(message.NewParentPath).Child(message.NewEntry.Name)
-		}
-
 		for _, sig := range message.Signatures {
 			if sig == targetFilerSignature && targetFilerSignature != 0 {
 				fmt.Printf("%s skipping %s change to %v\n", targetFiler, sourceFiler, message)
 				return nil
 			}
 		}
-		if debug {
-			fmt.Printf("%s check %s change %s,%s sig %v, target sig: %v\n", targetFiler, sourceFiler, sourceOldKey, sourceNewKey, message.Signatures, targetFilerSignature)
-		}
-
-		if !strings.HasPrefix(resp.Directory, sourcePath) {
-			return nil
-		}
-
-		// handle deletions
-		if message.OldEntry != nil && message.NewEntry == nil {
-			if !strings.HasPrefix(string(sourceOldKey), sourcePath) {
-				return nil
-			}
-			key := util.Join(targetPath, string(sourceOldKey)[len(sourcePath):])
-			return filerSink.DeleteEntry(key, message.OldEntry.IsDirectory, message.DeleteChunks, message.Signatures)
-		}
-
-		// handle new entries
-		if message.OldEntry == nil && message.NewEntry != nil {
-			if !strings.HasPrefix(string(sourceNewKey), sourcePath) {
-				return nil
-			}
-			key := util.Join(targetPath, string(sourceNewKey)[len(sourcePath):])
-			return filerSink.CreateEntry(key, message.NewEntry, message.Signatures)
-		}
-
-		// this is something special?
-		if message.OldEntry == nil && message.NewEntry == nil {
-			return nil
-		}
-
-		// handle updates
-		if strings.HasPrefix(string(sourceOldKey), sourcePath) {
-			// old key is in the watched directory
-			if strings.HasPrefix(string(sourceNewKey), sourcePath) {
-				// new key is also in the watched directory
-				oldKey := util.Join(targetPath, string(sourceOldKey)[len(sourcePath):])
-				message.NewParentPath = util.Join(targetPath, message.NewParentPath[len(sourcePath):])
-				foundExisting, err := filerSink.UpdateEntry(string(oldKey), message.OldEntry, message.NewParentPath, message.NewEntry, message.DeleteChunks, message.Signatures)
-				if foundExisting {
-					return err
-				}
-
-				// not able to find old entry
-				if err = filerSink.DeleteEntry(string(oldKey), message.OldEntry.IsDirectory, false, message.Signatures); err != nil {
-					return fmt.Errorf("delete old entry %v: %v", oldKey, err)
-				}
-
-				// create the new entry
-				newKey := util.Join(targetPath, string(sourceNewKey)[len(sourcePath):])
-				return filerSink.CreateEntry(newKey, message.NewEntry, message.Signatures)
-
-			} else {
-				// new key is outside of the watched directory
-				key := util.Join(targetPath, string(sourceOldKey)[len(sourcePath):])
-				return filerSink.DeleteEntry(key, message.OldEntry.IsDirectory, message.DeleteChunks, message.Signatures)
-			}
-		} else {
-			// old key is outside of the watched directory
-			if strings.HasPrefix(string(sourceNewKey), sourcePath) {
-				// new key is in the watched directory
-				key := util.Join(targetPath, string(sourceNewKey)[len(sourcePath):])
-				return filerSink.CreateEntry(key, message.NewEntry, message.Signatures)
-			} else {
-				// new key is also outside of the watched directory
-				// skip
-			}
-		}
-
-		return nil
+		return persistEventFn(resp)
 	}
 
 	return pb.WithFilerClient(sourceFiler, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
@@ -275,7 +200,7 @@ func doSubscribeFilerMetaChanges(grpcDialOption grpc.DialOption, sourceFiler, so
 				glog.V(0).Infof("sync %s => %s progressed to %v %0.2f/sec", sourceFiler, targetFiler, time.Unix(0, resp.TsNs), float64(counter)/float64(3))
 				counter = 0
 				lastWriteTime = time.Now()
-				if err := writeSyncOffset(grpcDialOption, targetFiler, sourceFilerSignature, resp.TsNs); err != nil {
+				if err := setOffset(grpcDialOption, targetFiler, SyncKeyPrefix, sourceFilerSignature, resp.TsNs); err != nil {
 					return err
 				}
 			}
@@ -290,11 +215,11 @@ const (
 	SyncKeyPrefix = "sync."
 )
 
-func readSyncOffset(grpcDialOption grpc.DialOption, filer string, filerSignature int32) (lastOffsetTsNs int64, readErr error) {
+func getOffset(grpcDialOption grpc.DialOption, filer string, signaturePrefix string, signature int32) (lastOffsetTsNs int64, readErr error) {
 
 	readErr = pb.WithFilerClient(filer, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
-		syncKey := []byte(SyncKeyPrefix + "____")
-		util.Uint32toBytes(syncKey[len(SyncKeyPrefix):len(SyncKeyPrefix)+4], uint32(filerSignature))
+		syncKey := []byte(signaturePrefix + "____")
+		util.Uint32toBytes(syncKey[len(signaturePrefix):len(signaturePrefix)+4], uint32(signature))
 
 		resp, err := client.KvGet(context.Background(), &filer_pb.KvGetRequest{Key: syncKey})
 		if err != nil {
@@ -317,11 +242,11 @@ func readSyncOffset(grpcDialOption grpc.DialOption, filer string, filerSignature
 
 }
 
-func writeSyncOffset(grpcDialOption grpc.DialOption, filer string, filerSignature int32, offsetTsNs int64) error {
+func setOffset(grpcDialOption grpc.DialOption, filer string, signaturePrefix string, signature int32, offsetTsNs int64) error {
 	return pb.WithFilerClient(filer, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
 
-		syncKey := []byte(SyncKeyPrefix + "____")
-		util.Uint32toBytes(syncKey[len(SyncKeyPrefix):len(SyncKeyPrefix)+4], uint32(filerSignature))
+		syncKey := []byte(signaturePrefix + "____")
+		util.Uint32toBytes(syncKey[len(signaturePrefix):len(signaturePrefix)+4], uint32(signature))
 
 		valueBuf := make([]byte, 8)
 		util.Uint64toBytes(valueBuf, uint64(offsetTsNs))
@@ -343,3 +268,107 @@ func writeSyncOffset(grpcDialOption grpc.DialOption, filer string, filerSignatur
 	})
 
 }
+
+func genProcessFunction(sourcePath string, targetPath string, dataSink sink.ReplicationSink, debug bool) func(resp *filer_pb.SubscribeMetadataResponse) error {
+	// process function
+	processEventFn := func(resp *filer_pb.SubscribeMetadataResponse) error {
+		message := resp.EventNotification
+
+		var sourceOldKey, sourceNewKey util.FullPath
+		if message.OldEntry != nil {
+			sourceOldKey = util.FullPath(resp.Directory).Child(message.OldEntry.Name)
+		}
+		if message.NewEntry != nil {
+			sourceNewKey = util.FullPath(message.NewParentPath).Child(message.NewEntry.Name)
+		}
+
+		if debug {
+			glog.V(0).Infof("received %v", resp)
+		}
+
+		if !strings.HasPrefix(resp.Directory, sourcePath) {
+			return nil
+		}
+
+		// handle deletions
+		if message.OldEntry != nil && message.NewEntry == nil {
+			if !strings.HasPrefix(string(sourceOldKey), sourcePath) {
+				return nil
+			}
+			key := buildKey(dataSink, message, targetPath, sourceOldKey, sourcePath)
+			return dataSink.DeleteEntry(key, message.OldEntry.IsDirectory, message.DeleteChunks, message.Signatures)
+		}
+
+		// handle new entries
+		if message.OldEntry == nil && message.NewEntry != nil {
+			if !strings.HasPrefix(string(sourceNewKey), sourcePath) {
+				return nil
+			}
+			key := buildKey(dataSink, message, targetPath, sourceNewKey, sourcePath)
+			return dataSink.CreateEntry(key, message.NewEntry, message.Signatures)
+		}
+
+		// this is something special?
+		if message.OldEntry == nil && message.NewEntry == nil {
+			return nil
+		}
+
+		// handle updates
+		if strings.HasPrefix(string(sourceOldKey), sourcePath) {
+			// old key is in the watched directory
+			if strings.HasPrefix(string(sourceNewKey), sourcePath) {
+				// new key is also in the watched directory
+				if !dataSink.IsIncremental() {
+					oldKey := util.Join(targetPath, string(sourceOldKey)[len(sourcePath):])
+					message.NewParentPath = util.Join(targetPath, message.NewParentPath[len(sourcePath):])
+					foundExisting, err := dataSink.UpdateEntry(string(oldKey), message.OldEntry, message.NewParentPath, message.NewEntry, message.DeleteChunks, message.Signatures)
+					if foundExisting {
+						return err
+					}
+
+					// not able to find old entry
+					if err = dataSink.DeleteEntry(string(oldKey), message.OldEntry.IsDirectory, false, message.Signatures); err != nil {
+						return fmt.Errorf("delete old entry %v: %v", oldKey, err)
+					}
+				}
+				// create the new entry
+				newKey := buildKey(dataSink, message, targetPath, sourceNewKey, sourcePath)
+				return dataSink.CreateEntry(newKey, message.NewEntry, message.Signatures)
+
+			} else {
+				// new key is outside of the watched directory
+				if !dataSink.IsIncremental() {
+					key := buildKey(dataSink, message, targetPath, sourceOldKey, sourcePath)
+					return dataSink.DeleteEntry(key, message.OldEntry.IsDirectory, message.DeleteChunks, message.Signatures)
+				}
+			}
+		} else {
+			// old key is outside of the watched directory
+			if strings.HasPrefix(string(sourceNewKey), sourcePath) {
+				// new key is in the watched directory
+				key := buildKey(dataSink, message, targetPath, sourceNewKey, sourcePath)
+				return dataSink.CreateEntry(key, message.NewEntry, message.Signatures)
+			} else {
+				// new key is also outside of the watched directory
+				// skip
+			}
+		}
+
+		return nil
+	}
+	return processEventFn
+}
+
+func buildKey(dataSink sink.ReplicationSink, message *filer_pb.EventNotification, targetPath string, sourceKey util.FullPath, sourcePath string) string {
+	if !dataSink.IsIncremental() {
+		return util.Join(targetPath, string(sourceKey)[len(sourcePath):])
+	}
+	var mTime int64
+	if message.NewEntry != nil {
+		mTime = message.NewEntry.Attributes.Mtime
+	} else if message.OldEntry != nil {
+		mTime = message.OldEntry.Attributes.Mtime
+	}
+	dateKey := time.Unix(mTime, 0).Format("2006-01-02")
+	return util.Join(targetPath, dateKey, string(sourceKey)[len(sourcePath):])
+}

weed/command/master.go

@@ -6,7 +6,6 @@ import (
 	"google.golang.org/grpc/reflection"
 	"net/http"
 	"os"
-	"runtime"
 	"sort"
 	"strconv"
 	"strings"
@@ -48,8 +47,8 @@ type MasterOptions struct {
 func init() {
 	cmdMaster.Run = runMaster // break init cycle
 	m.port = cmdMaster.Flag.Int("port", 9333, "http listen port")
-	m.ip = cmdMaster.Flag.String("ip", util.DetectedHostAddress(), "master <ip>|<server> address")
-	m.ipBind = cmdMaster.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to")
+	m.ip = cmdMaster.Flag.String("ip", util.DetectedHostAddress(), "master <ip>|<server> address, also used as identifier")
+	m.ipBind = cmdMaster.Flag.String("ip.bind", "", "ip address to bind to")
 	m.metaFolder = cmdMaster.Flag.String("mdir", os.TempDir(), "data directory to store meta data")
 	m.peers = cmdMaster.Flag.String("peers", "", "all master nodes in comma separated ip:port list, example: 127.0.0.1:9093,127.0.0.1:9094,127.0.0.1:9095")
 	m.volumeSizeLimitMB = cmdMaster.Flag.Uint("volumeSizeLimitMB", 30*1000, "Master stops directing writes to oversized volumes.")
@@ -86,7 +85,6 @@ func runMaster(cmd *Command, args []string) bool {
 	util.LoadConfiguration("security", false)
 	util.LoadConfiguration("master", false)
 
-	runtime.GOMAXPROCS(runtime.NumCPU())
 	grace.SetupProfiling(*masterCpuProfile, *masterMemProfile)
 
 	parent, _ := util.FullPath(*m.metaFolder).DirAndName()
@@ -138,7 +136,6 @@ func startMaster(masterOption MasterOptions, masterWhiteList []string) {
 	if err != nil {
 		glog.Fatalf("master failed to listen on grpc port %d: %v", grpcPort, err)
 	}
-	// Create your protocol servers.
 	grpcS := pb.NewGrpcServer(security.LoadServerTLS(util.GetViper(), "grpc.master"))
 	master_pb.RegisterSeaweedServer(grpcS, ms)
 	protobuf.RegisterRaftServer(grpcS, raftServer)

weed/command/mount.go

@@ -25,6 +25,7 @@ type MountOptions struct {
 	volumeServerAccess *string
 	uidMap             *string
 	gidMap             *string
+	readOnly           *bool
 }
 
 var (
@@ -42,10 +43,10 @@ func init() {
 	mountOptions.dirAutoCreate = cmdMount.Flag.Bool("dirAutoCreate", false, "auto create the directory to mount to")
 	mountOptions.collection = cmdMount.Flag.String("collection", "", "collection to create the files")
 	mountOptions.replication = cmdMount.Flag.String("replication", "", "replication(e.g. 000, 001) to create to files. If empty, let filer decide.")
-	mountOptions.diskType = cmdMount.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
+	mountOptions.diskType = cmdMount.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	mountOptions.ttlSec = cmdMount.Flag.Int("ttl", 0, "file ttl in seconds")
 	mountOptions.chunkSizeLimitMB = cmdMount.Flag.Int("chunkSizeLimitMB", 2, "local write buffer size, also chunk large files")
-	mountOptions.concurrentWriters = cmdMount.Flag.Int("concurrentWriters", 128, "limit concurrent goroutine writers if not 0")
+	mountOptions.concurrentWriters = cmdMount.Flag.Int("concurrentWriters", 32, "limit concurrent goroutine writers if not 0")
 	mountOptions.cacheDir = cmdMount.Flag.String("cacheDir", os.TempDir(), "local cache directory for file chunks and meta data")
 	mountOptions.cacheSizeMB = cmdMount.Flag.Int64("cacheCapacityMB", 1000, "local file chunk cache capacity in MB (0 will disable cache)")
 	mountOptions.dataCenter = cmdMount.Flag.String("dataCenter", "", "prefer to write to the data center")
@@ -55,6 +56,7 @@ func init() {
 	mountOptions.volumeServerAccess = cmdMount.Flag.String("volumeServerAccess", "direct", "access volume servers by [direct|publicUrl|filerProxy]")
 	mountOptions.uidMap = cmdMount.Flag.String("map.uid", "", "map local uid to uid on filer, comma-separated <local_uid>:<filer_uid>")
 	mountOptions.gidMap = cmdMount.Flag.String("map.gid", "", "map local gid to gid on filer, comma-separated <local_gid>:<filer_gid>")
+	mountOptions.readOnly = cmdMount.Flag.Bool("readOnly", false, "read only")
 
 	mountCpuProfile = cmdMount.Flag.String("cpuprofile", "", "cpu profile output file")
 	mountMemProfile = cmdMount.Flag.String("memprofile", "", "memory profile output file")

weed/command/mount_std.go

@@ -53,7 +53,7 @@ func RunMount(option *MountOptions, umask os.FileMode) bool {
 
 	filer := *option.filer
 	// parse filer grpc address
-	filerGrpcAddress, err := pb.ParseFilerGrpcAddress(filer)
+	filerGrpcAddress, err := pb.ParseServerToGrpcAddress(filer)
 	if err != nil {
 		glog.V(0).Infof("ParseFilerGrpcAddress: %v", err)
 		return true
@@ -63,16 +63,23 @@ func RunMount(option *MountOptions, umask os.FileMode) bool {
 	// try to connect to filer, filerBucketsPath may be useful later
 	grpcDialOption := security.LoadClientTLS(util.GetViper(), "grpc.client")
 	var cipher bool
-	err = pb.WithGrpcFilerClient(filerGrpcAddress, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
-		resp, err := client.GetFilerConfiguration(context.Background(), &filer_pb.GetFilerConfigurationRequest{})
+	for i := 0; i < 10; i++ {
+		err = pb.WithGrpcFilerClient(filerGrpcAddress, grpcDialOption, func(client filer_pb.SeaweedFilerClient) error {
+			resp, err := client.GetFilerConfiguration(context.Background(), &filer_pb.GetFilerConfigurationRequest{})
+			if err != nil {
+				return fmt.Errorf("get filer grpc address %s configuration: %v", filerGrpcAddress, err)
+			}
+			cipher = resp.Cipher
+			return nil
+		})
 		if err != nil {
-			return fmt.Errorf("get filer grpc address %s configuration: %v", filerGrpcAddress, err)
+			glog.V(0).Infof("failed to talk to filer %s: %v", filerGrpcAddress, err)
+			glog.V(0).Infof("wait for %d seconds ...", i+1)
+			time.Sleep(time.Duration(i+1) * time.Second)
 		}
-		cipher = resp.Cipher
-		return nil
-	})
+	}
 	if err != nil {
-		glog.Infof("failed to talk to filer %s: %v", filerGrpcAddress, err)
+		glog.Errorf("failed to talk to filer %s: %v", filerGrpcAddress, err)
 		return true
 	}
 
@@ -196,6 +203,7 @@ func RunMount(option *MountOptions, umask os.FileMode) bool {
 		VolumeServerAccess: *mountOptions.volumeServerAccess,
 		Cipher:             cipher,
 		UidGidMapper:       uidGidMapper,
+		ReadOnly:           *option.readOnly,
 	})
 
 	// mount

weed/command/msg_broker.go

@@ -63,7 +63,7 @@ func (msgBrokerOpt *MessageBrokerOptions) startQueueServer() bool {
 
 	grace.SetupProfiling(*messageBrokerStandaloneOptions.cpuprofile, *messageBrokerStandaloneOptions.memprofile)
 
-	filerGrpcAddress, err := pb.ParseFilerGrpcAddress(*msgBrokerOpt.filer)
+	filerGrpcAddress, err := pb.ParseServerToGrpcAddress(*msgBrokerOpt.filer)
 	if err != nil {
 		glog.Fatal(err)
 		return false

weed/command/s3.go

@@ -137,7 +137,7 @@ func runS3(cmd *Command, args []string) bool {
 
 func (s3opt *S3Options) startS3Server() bool {
 
-	filerGrpcAddress, err := pb.ParseFilerGrpcAddress(*s3opt.filer)
+	filerGrpcAddress, err := pb.ParseServerToGrpcAddress(*s3opt.filer)
 	if err != nil {
 		glog.Fatal(err)
 		return false

weed/command/scaffold.go

@@ -103,9 +103,9 @@ dir = "./filerrdb"					# directory to store rocksdb files
 
 [mysql]  # or memsql, tidb
 # CREATE TABLE IF NOT EXISTS filemeta (
-#   dirhash     BIGINT         COMMENT 'first 64 bits of MD5 hash value of directory field',
-#   name        VARCHAR(1000)  COMMENT 'directory or file name',
-#   directory   TEXT           COMMENT 'full path to parent directory',
+#   dirhash     BIGINT               COMMENT 'first 64 bits of MD5 hash value of directory field',
+#   name        VARCHAR(1000) BINARY COMMENT 'directory or file name',
+#   directory   TEXT                 COMMENT 'full path to parent directory',
 #   meta        LONGBLOB,
 #   PRIMARY KEY (dirhash, name)
 # ) DEFAULT CHARSET=utf8;
@@ -120,13 +120,16 @@ connection_max_idle = 2
 connection_max_open = 100
 connection_max_lifetime_seconds = 0
 interpolateParams = false
+# if insert/upsert failing, you can disable upsert or update query syntax to match your RDBMS syntax:
+enableUpsert = true
+upsertQuery = """INSERT INTO ` + "`%s`" + ` (dirhash,name,directory,meta) VALUES(?,?,?,?) ON DUPLICATE KEY UPDATE meta = VALUES(meta)"""
 
 [mysql2]  # or memsql, tidb
 enabled = false
 createTable = """
   CREATE TABLE IF NOT EXISTS ` + "`%s`" + ` (
     dirhash BIGINT,
-    name VARCHAR(1000),
+    name VARCHAR(1000) BINARY,
     directory TEXT,
     meta LONGBLOB,
     PRIMARY KEY (dirhash, name)
@@ -141,6 +144,9 @@ connection_max_idle = 2
 connection_max_open = 100
 connection_max_lifetime_seconds = 0
 interpolateParams = false
+# if insert/upsert failing, you can disable upsert or update query syntax to match your RDBMS syntax:
+enableUpsert = true
+upsertQuery = """INSERT INTO ` + "`%s`" + ` (dirhash,name,directory,meta) VALUES(?,?,?,?) ON DUPLICATE KEY UPDATE meta = VALUES(meta)"""
 
 [postgres] # or cockroachdb, YugabyteDB
 # CREATE TABLE IF NOT EXISTS filemeta (
@@ -161,6 +167,9 @@ sslmode = "disable"
 connection_max_idle = 100
 connection_max_open = 100
 connection_max_lifetime_seconds = 0
+# if insert/upsert failing, you can disable upsert or update query syntax to match your RDBMS syntax:
+enableUpsert = true
+upsertQuery = """INSERT INTO "%[1]s" (dirhash,name,directory,meta) VALUES($1,$2,$3,$4) ON CONFLICT (dirhash,name) DO UPDATE SET meta = EXCLUDED.meta WHERE "%[1]s".meta != EXCLUDED.meta"""
 
 [postgres2]
 enabled = false
@@ -183,6 +192,9 @@ sslmode = "disable"
 connection_max_idle = 100
 connection_max_open = 100
 connection_max_lifetime_seconds = 0
+# if insert/upsert failing, you can disable upsert or update query syntax to match your RDBMS syntax:
+enableUpsert = true
+upsertQuery = """INSERT INTO "%[1]s" (dirhash,name,directory,meta) VALUES($1,$2,$3,$4) ON CONFLICT (dirhash,name) DO UPDATE SET meta = EXCLUDED.meta WHERE "%[1]s".meta != EXCLUDED.meta"""
 
 [cassandra]
 # CREATE TABLE filemeta (
@@ -356,6 +368,9 @@ directory = "/buckets"
 [sink.local]
 enabled = false
 directory = "/data"
+# all replicated files are under modified time as yyyy-mm-dd directories
+# so each date directory contains all new and updated files.
+is_incremental = false
 
 [sink.local_incremental]
 # all replicated files are under modified time as yyyy-mm-dd directories
@@ -373,6 +388,7 @@ directory = "/backup"
 replication = ""
 collection = ""
 ttlSec = 0
+is_incremental = false
 
 [sink.s3]
 # read credentials doc at https://docs.aws.amazon.com/sdk-for-go/v1/developer-guide/sessions.html
@@ -384,6 +400,7 @@ region = "us-east-2"
 bucket = "your_bucket_name"    # an existing bucket
 directory = "/"                # destination directory
 endpoint = ""
+is_incremental = false
 
 [sink.google_cloud_storage]
 # read credentials doc at https://cloud.google.com/docs/authentication/getting-started
@@ -391,6 +408,7 @@ enabled = false
 google_application_credentials = "/path/to/x.json" # path to json credential file
 bucket = "your_bucket_seaweedfs"    # an existing bucket
 directory = "/"                     # destination directory
+is_incremental = false
 
 [sink.azure]
 # experimental, let me know if it works
@@ -399,6 +417,7 @@ account_name = ""
 account_key  = ""
 container = "mycontainer"      # an existing container
 directory = "/"                # destination directory
+is_incremental = false
 
 [sink.backblaze]
 enabled = false
@@ -406,6 +425,7 @@ b2_account_id = ""
 b2_master_application_key  = ""
 bucket = "mybucket"            # an existing bucket
 directory = "/"                # destination directory
+is_incremental = false
 
 `
 
@@ -432,22 +452,28 @@ expires_after_seconds = 10           # seconds
 # the host name is not checked, so the PERM files can be shared.
 [grpc]
 ca = ""
+# Set wildcard domain for enable TLS authentication by common names
+allowed_wildcard_domain = "" # .mycompany.com
 
 [grpc.volume]
 cert = ""
 key  = ""
+allowed_commonNames = ""	# comma-separated SSL certificate common names
 
 [grpc.master]
 cert = ""
 key  = ""
+allowed_commonNames = ""	# comma-separated SSL certificate common names
 
 [grpc.filer]
 cert = ""
 key  = ""
+allowed_commonNames = ""	# comma-separated SSL certificate common names
 
 [grpc.msg_broker]
 cert = ""
 key  = ""
+allowed_commonNames = ""	# comma-separated SSL certificate common names
 
 # use this for any place needs a grpc client
 # i.e., "weed backup|benchmark|filer.copy|filer.replicate|mount|s3|upload"
@@ -455,7 +481,6 @@ key  = ""
 cert = ""
 key  = ""
 
-
 # volume server https options
 # Note: work in progress!
 #     this does not work with other clients, e.g., "weed filer|mount" etc, yet.
@@ -493,7 +518,7 @@ default = "localhost:8888"    # used by maintenance scripts if the scripts needs
 
 
 [master.sequencer]
-type = "raft"     # Choose [raft|etcd] type for storing the file id sequence
+type = "raft"     # Choose [raft|etcd|snowflake] type for storing the file id sequence
 # when sequencer.type = etcd, set listen client urls of etcd cluster that store file id sequence
 # example : http://127.0.0.1:2379,http://127.0.0.1:2389
 sequencer_etcd_urls = "http://127.0.0.1:2379"

weed/command/server.go

@@ -2,9 +2,8 @@ package command
 
 import (
 	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/util/grace"
 	"os"
-	"runtime"
-	"runtime/pprof"
 	"strings"
 	"time"
 
@@ -16,6 +15,7 @@ import (
 
 type ServerOptions struct {
 	cpuprofile *string
+	memprofile *string
 	v          VolumeServerOptions
 }
 
@@ -49,8 +49,8 @@ var cmdServer = &Command{
 }
 
 var (
-	serverIp                  = cmdServer.Flag.String("ip", util.DetectedHostAddress(), "ip or server name")
-	serverBindIp              = cmdServer.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to")
+	serverIp                  = cmdServer.Flag.String("ip", util.DetectedHostAddress(), "ip or server name, also used as identifier")
+	serverBindIp              = cmdServer.Flag.String("ip.bind", "", "ip address to bind to")
 	serverTimeout             = cmdServer.Flag.Int("idleTimeout", 30, "connection idle seconds")
 	serverDataCenter          = cmdServer.Flag.String("dataCenter", "", "current volume server's data center name")
 	serverRack                = cmdServer.Flag.String("rack", "", "current volume server's rack name")
@@ -76,6 +76,7 @@ var (
 
 func init() {
 	serverOptions.cpuprofile = cmdServer.Flag.String("cpuprofile", "", "cpu profile output file")
+	serverOptions.memprofile = cmdServer.Flag.String("memprofile", "", "memory profile output file")
 
 	masterOptions.port = cmdServer.Flag.Int("master.port", 9333, "master server http listen port")
 	masterOptions.metaFolder = cmdServer.Flag.String("master.dir", "", "data directory to store meta data, default to same as -dir specified")
@@ -93,24 +94,27 @@ func init() {
 	filerOptions.publicPort = cmdServer.Flag.Int("filer.port.public", 0, "filer server public http listen port")
 	filerOptions.defaultReplicaPlacement = cmdServer.Flag.String("filer.defaultReplicaPlacement", "", "default replication type. If not specified, use master setting.")
 	filerOptions.disableDirListing = cmdServer.Flag.Bool("filer.disableDirListing", false, "turn off directory listing")
-	filerOptions.maxMB = cmdServer.Flag.Int("filer.maxMB", 32, "split files larger than the limit")
+	filerOptions.maxMB = cmdServer.Flag.Int("filer.maxMB", 4, "split files larger than the limit")
 	filerOptions.dirListingLimit = cmdServer.Flag.Int("filer.dirListLimit", 1000, "limit sub dir listing size")
 	filerOptions.cipher = cmdServer.Flag.Bool("filer.encryptVolumeData", false, "encrypt data on volume servers")
 	filerOptions.peers = cmdServer.Flag.String("filer.peers", "", "all filers sharing the same filer store in comma separated ip:port list")
 	filerOptions.saveToFilerLimit = cmdServer.Flag.Int("filer.saveToFilerLimit", 0, "Small files smaller than this limit can be cached in filer store.")
+	filerOptions.concurrentUploadLimitMB = cmdServer.Flag.Int("filer.concurrentUploadLimitMB", 64, "limit total concurrent upload size")
 
 	serverOptions.v.port = cmdServer.Flag.Int("volume.port", 8080, "volume server http listen port")
 	serverOptions.v.publicPort = cmdServer.Flag.Int("volume.port.public", 0, "volume server public port")
 	serverOptions.v.indexType = cmdServer.Flag.String("volume.index", "memory", "Choose [memory|leveldb|leveldbMedium|leveldbLarge] mode for memory~performance balance.")
-	serverOptions.v.diskType = cmdServer.Flag.String("volume.disk", "", "[hdd|ssd] hard drive or solid state drive")
+	serverOptions.v.diskType = cmdServer.Flag.String("volume.disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	serverOptions.v.fixJpgOrientation = cmdServer.Flag.Bool("volume.images.fix.orientation", false, "Adjust jpg orientation when uploading.")
 	serverOptions.v.readRedirect = cmdServer.Flag.Bool("volume.read.redirect", true, "Redirect moved or non-local volumes.")
 	serverOptions.v.compactionMBPerSecond = cmdServer.Flag.Int("volume.compactionMBps", 0, "limit compaction speed in mega bytes per second")
 	serverOptions.v.fileSizeLimitMB = cmdServer.Flag.Int("volume.fileSizeLimitMB", 256, "limit file size to avoid out of memory")
+	serverOptions.v.concurrentUploadLimitMB = cmdServer.Flag.Int("volume.concurrentUploadLimitMB", 64, "limit total concurrent upload size")
 	serverOptions.v.publicUrl = cmdServer.Flag.String("volume.publicUrl", "", "publicly accessible address")
 	serverOptions.v.preStopSeconds = cmdServer.Flag.Int("volume.preStopSeconds", 10, "number of seconds between stop send heartbeats and stop volume server")
 	serverOptions.v.pprof = cmdServer.Flag.Bool("volume.pprof", false, "enable pprof http handlers. precludes --memprofile and --cpuprofile")
 	serverOptions.v.idxFolder = cmdServer.Flag.String("volume.dir.idx", "", "directory to store .idx files")
+	serverOptions.v.enableTcp = cmdServer.Flag.Bool("volume.tcp", false, "<exprimental> enable tcp port")
 
 	s3Options.port = cmdServer.Flag.Int("s3.port", 8333, "s3 server http listen port")
 	s3Options.domainName = cmdServer.Flag.String("s3.domainName", "", "suffix of the host name in comma separated list, {bucket}.{domainName}")
@@ -122,7 +126,7 @@ func init() {
 	webdavOptions.port = cmdServer.Flag.Int("webdav.port", 7333, "webdav server http listen port")
 	webdavOptions.collection = cmdServer.Flag.String("webdav.collection", "", "collection to create the files")
 	webdavOptions.replication = cmdServer.Flag.String("webdav.replication", "", "replication to create the files")
-	webdavOptions.disk = cmdServer.Flag.String("webdav.disk", "", "[hdd|ssd] hard drive or solid state drive")
+	webdavOptions.disk = cmdServer.Flag.String("webdav.disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	webdavOptions.tlsPrivateKey = cmdServer.Flag.String("webdav.key.file", "", "path to the TLS private key file")
 	webdavOptions.tlsCertificate = cmdServer.Flag.String("webdav.cert.file", "", "path to the TLS certificate file")
 	webdavOptions.cacheDir = cmdServer.Flag.String("webdav.cacheDir", os.TempDir(), "local cache directory for file chunks")
@@ -137,14 +141,7 @@ func runServer(cmd *Command, args []string) bool {
 	util.LoadConfiguration("security", false)
 	util.LoadConfiguration("master", false)
 
-	if *serverOptions.cpuprofile != "" {
-		f, err := os.Create(*serverOptions.cpuprofile)
-		if err != nil {
-			glog.Fatal(err)
-		}
-		pprof.StartCPUProfile(f)
-		defer pprof.StopCPUProfile()
-	}
+	grace.SetupProfiling(*serverOptions.cpuprofile, *serverOptions.memprofile)
 
 	if *isStartingS3 {
 		*isStartingFiler = true
@@ -156,19 +153,21 @@ func runServer(cmd *Command, args []string) bool {
 		*isStartingFiler = true
 	}
 
-	_, peerList := checkPeers(*serverIp, *masterOptions.port, *masterOptions.peers)
-	peers := strings.Join(peerList, ",")
-	masterOptions.peers = &peers
+	if *isStartingMasterServer {
+		_, peerList := checkPeers(*serverIp, *masterOptions.port, *masterOptions.peers)
+		peers := strings.Join(peerList, ",")
+		masterOptions.peers = &peers
+	}
 
 	// ip address
 	masterOptions.ip = serverIp
 	masterOptions.ipBind = serverBindIp
-	filerOptions.masters = &peers
+	filerOptions.masters = masterOptions.peers
 	filerOptions.ip = serverIp
 	filerOptions.bindIp = serverBindIp
 	serverOptions.v.ip = serverIp
 	serverOptions.v.bindIp = serverBindIp
-	serverOptions.v.masters = &peers
+	serverOptions.v.masters = masterOptions.peers
 	serverOptions.v.idleConnectionTimeout = serverTimeout
 	serverOptions.v.dataCenter = serverDataCenter
 	serverOptions.v.rack = serverRack
@@ -189,7 +188,6 @@ func runServer(cmd *Command, args []string) bool {
 	webdavOptions.filer = &filerAddress
 	msgBrokerOptions.filer = &filerAddress
 
-	runtime.GOMAXPROCS(runtime.NumCPU())
 	go stats_collect.StartMetricsServer(*serverMetricsHttpPort)
 
 	folders := strings.Split(*volumeDataFolders, ",")

weed/command/upload.go

@@ -41,9 +41,9 @@ func init() {
 	upload.replication = cmdUpload.Flag.String("replication", "", "replication type")
 	upload.collection = cmdUpload.Flag.String("collection", "", "optional collection name")
 	upload.dataCenter = cmdUpload.Flag.String("dataCenter", "", "optional data center name")
-	upload.diskType = cmdUpload.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
+	upload.diskType = cmdUpload.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	upload.ttl = cmdUpload.Flag.String("ttl", "", "time to live, e.g.: 1m, 1h, 1d, 1M, 1y")
-	upload.maxMB = cmdUpload.Flag.Int("maxMB", 32, "split files larger than the limit")
+	upload.maxMB = cmdUpload.Flag.Int("maxMB", 4, "split files larger than the limit")
 	upload.usePublicUrl = cmdUpload.Flag.Bool("usePublicUrl", false, "upload to public url from volume server")
 }
 

weed/command/volume.go

@@ -6,7 +6,6 @@ import (
 	"net/http"
 	httppprof "net/http/pprof"
 	"os"
-	"runtime"
 	"runtime/pprof"
 	"strconv"
 	"strings"
@@ -36,41 +35,43 @@ var (
 )
 
 type VolumeServerOptions struct {
-	port                  *int
-	publicPort            *int
-	folders               []string
-	folderMaxLimits       []int
-	idxFolder             *string
-	ip                    *string
-	publicUrl             *string
-	bindIp                *string
-	masters               *string
-	idleConnectionTimeout *int
-	dataCenter            *string
-	rack                  *string
-	whiteList             []string
-	indexType             *string
-	diskType              *string
-	fixJpgOrientation     *bool
-	readRedirect          *bool
-	cpuProfile            *string
-	memProfile            *string
-	compactionMBPerSecond *int
-	fileSizeLimitMB       *int
-	minFreeSpacePercents  []float32
-	pprof                 *bool
-	preStopSeconds        *int
-	metricsHttpPort       *int
+	port                    *int
+	publicPort              *int
+	folders                 []string
+	folderMaxLimits         []int
+	idxFolder               *string
+	ip                      *string
+	publicUrl               *string
+	bindIp                  *string
+	masters                 *string
+	idleConnectionTimeout   *int
+	dataCenter              *string
+	rack                    *string
+	whiteList               []string
+	indexType               *string
+	diskType                *string
+	fixJpgOrientation       *bool
+	readRedirect            *bool
+	cpuProfile              *string
+	memProfile              *string
+	compactionMBPerSecond   *int
+	fileSizeLimitMB         *int
+	concurrentUploadLimitMB *int
+	minFreeSpacePercents    []float32
+	pprof                   *bool
+	preStopSeconds          *int
+	metricsHttpPort         *int
 	// pulseSeconds          *int
+	enableTcp *bool
 }
 
 func init() {
 	cmdVolume.Run = runVolume // break init cycle
 	v.port = cmdVolume.Flag.Int("port", 8080, "http listen port")
 	v.publicPort = cmdVolume.Flag.Int("port.public", 0, "port opened to public")
-	v.ip = cmdVolume.Flag.String("ip", util.DetectedHostAddress(), "ip or server name")
+	v.ip = cmdVolume.Flag.String("ip", util.DetectedHostAddress(), "ip or server name, also used as identifier")
 	v.publicUrl = cmdVolume.Flag.String("publicUrl", "", "Publicly accessible address")
-	v.bindIp = cmdVolume.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to")
+	v.bindIp = cmdVolume.Flag.String("ip.bind", "", "ip address to bind to")
 	v.masters = cmdVolume.Flag.String("mserver", "localhost:9333", "comma-separated master servers")
 	v.preStopSeconds = cmdVolume.Flag.Int("preStopSeconds", 10, "number of seconds between stop send heartbeats and stop volume server")
 	// v.pulseSeconds = cmdVolume.Flag.Int("pulseSeconds", 5, "number of seconds between heartbeats, must be smaller than or equal to the master's setting")
@@ -78,16 +79,18 @@ func init() {
 	v.dataCenter = cmdVolume.Flag.String("dataCenter", "", "current volume server's data center name")
 	v.rack = cmdVolume.Flag.String("rack", "", "current volume server's rack name")
 	v.indexType = cmdVolume.Flag.String("index", "memory", "Choose [memory|leveldb|leveldbMedium|leveldbLarge] mode for memory~performance balance.")
-	v.diskType = cmdVolume.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
+	v.diskType = cmdVolume.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	v.fixJpgOrientation = cmdVolume.Flag.Bool("images.fix.orientation", false, "Adjust jpg orientation when uploading.")
 	v.readRedirect = cmdVolume.Flag.Bool("read.redirect", true, "Redirect moved or non-local volumes.")
 	v.cpuProfile = cmdVolume.Flag.String("cpuprofile", "", "cpu profile output file")
 	v.memProfile = cmdVolume.Flag.String("memprofile", "", "memory profile output file")
 	v.compactionMBPerSecond = cmdVolume.Flag.Int("compactionMBps", 0, "limit background compaction or copying speed in mega bytes per second")
 	v.fileSizeLimitMB = cmdVolume.Flag.Int("fileSizeLimitMB", 256, "limit file size to avoid out of memory")
+	v.concurrentUploadLimitMB = cmdVolume.Flag.Int("concurrentUploadLimitMB", 128, "limit total concurrent upload size")
 	v.pprof = cmdVolume.Flag.Bool("pprof", false, "enable pprof http handlers. precludes --memprofile and --cpuprofile")
 	v.metricsHttpPort = cmdVolume.Flag.Int("metricsPort", 0, "Prometheus metrics listen port")
 	v.idxFolder = cmdVolume.Flag.String("dir.idx", "", "directory to store .idx files")
+	v.enableTcp = cmdVolume.Flag.Bool("tcp", false, "<exprimental> enable tcp port")
 }
 
 var cmdVolume = &Command{
@@ -109,8 +112,6 @@ func runVolume(cmd *Command, args []string) bool {
 
 	util.LoadConfiguration("security", false)
 
-	runtime.GOMAXPROCS(runtime.NumCPU())
-
 	// If --pprof is set we assume the caller wants to be able to collect
 	// cpu and memory profiles via go tool pprof
 	if !*v.pprof {
@@ -238,6 +239,7 @@ func (v VolumeServerOptions) startVolumeServer(volumeFolders, maxVolumeCounts, v
 		*v.fixJpgOrientation, *v.readRedirect,
 		*v.compactionMBPerSecond,
 		*v.fileSizeLimitMB,
+		int64(*v.concurrentUploadLimitMB)*1024*1024,
 	)
 	// starting grpc server
 	grpcS := v.startGrpcService(volumeServer)
@@ -251,6 +253,11 @@ func (v VolumeServerOptions) startVolumeServer(volumeFolders, maxVolumeCounts, v
 		}
 	}
 
+	// starting tcp server
+	if *v.enableTcp {
+		go v.startTcpService(volumeServer)
+	}
+
 	// starting the cluster http server
 	clusterHttpServer := v.startClusterHttpService(volumeMux)
 
@@ -368,3 +375,22 @@ func (v VolumeServerOptions) startClusterHttpService(handler http.Handler) httpd
 	}()
 	return clusterHttpServer
 }
+
+func (v VolumeServerOptions) startTcpService(volumeServer *weed_server.VolumeServer) {
+	listeningAddress := *v.bindIp + ":" + strconv.Itoa(*v.port+20000)
+	glog.V(0).Infoln("Start Seaweed volume server", util.Version(), "tcp at", listeningAddress)
+	listener, e := util.NewListener(listeningAddress, 0)
+	if e != nil {
+		glog.Fatalf("Volume server listener error on %s:%v", listeningAddress, e)
+	}
+	defer listener.Close()
+
+	for {
+		c, err := listener.Accept()
+		if err != nil {
+			fmt.Println(err)
+			return
+		}
+		go volumeServer.HandleTcpConnection(c)
+	}
+}

weed/command/webdav.go

@@ -39,7 +39,7 @@ func init() {
 	webDavStandaloneOptions.port = cmdWebDav.Flag.Int("port", 7333, "webdav server http listen port")
 	webDavStandaloneOptions.collection = cmdWebDav.Flag.String("collection", "", "collection to create the files")
 	webDavStandaloneOptions.replication = cmdWebDav.Flag.String("replication", "", "replication to create the files")
-	webDavStandaloneOptions.disk = cmdWebDav.Flag.String("disk", "", "[hdd|ssd] hard drive or solid state drive")
+	webDavStandaloneOptions.disk = cmdWebDav.Flag.String("disk", "", "[hdd|ssd|<tag>] hard drive or solid state drive or any tag")
 	webDavStandaloneOptions.tlsPrivateKey = cmdWebDav.Flag.String("key.file", "", "path to the TLS private key file")
 	webDavStandaloneOptions.tlsCertificate = cmdWebDav.Flag.String("cert.file", "", "path to the TLS certificate file")
 	webDavStandaloneOptions.cacheDir = cmdWebDav.Flag.String("cacheDir", os.TempDir(), "local cache directory for file chunks")
@@ -78,7 +78,7 @@ func (wo *WebDavOption) startWebDav() bool {
 	}
 
 	// parse filer grpc address
-	filerGrpcAddress, err := pb.ParseFilerGrpcAddress(*wo.filer)
+	filerGrpcAddress, err := pb.ParseServerToGrpcAddress(*wo.filer)
 	if err != nil {
 		glog.Fatal(err)
 		return false

weed/filer/abstract_sql/abstract_sql_store.go

@@ -13,15 +13,15 @@ import (
 )
 
 type SqlGenerator interface {
-	GetSqlInsert(bucket string) string
-	GetSqlUpdate(bucket string) string
-	GetSqlFind(bucket string) string
-	GetSqlDelete(bucket string) string
-	GetSqlDeleteFolderChildren(bucket string) string
-	GetSqlListExclusive(bucket string) string
-	GetSqlListInclusive(bucket string) string
-	GetSqlCreateTable(bucket string) string
-	GetSqlDropTable(bucket string) string
+	GetSqlInsert(tableName string) string
+	GetSqlUpdate(tableName string) string
+	GetSqlFind(tableName string) string
+	GetSqlDelete(tableName string) string
+	GetSqlDeleteFolderChildren(tableName string) string
+	GetSqlListExclusive(tableName string) string
+	GetSqlListInclusive(tableName string) string
+	GetSqlCreateTable(tableName string) string
+	GetSqlDropTable(tableName string) string
 }
 
 type AbstractSqlStore struct {
@@ -32,6 +32,29 @@ type AbstractSqlStore struct {
 	dbsLock            sync.Mutex
 }
 
+func (store *AbstractSqlStore) OnBucketCreation(bucket string) {
+	store.dbsLock.Lock()
+	defer store.dbsLock.Unlock()
+
+	store.CreateTable(context.Background(), bucket)
+
+	if store.dbs == nil {
+		return
+	}
+	store.dbs[bucket] = true
+}
+func (store *AbstractSqlStore) OnBucketDeletion(bucket string) {
+	store.dbsLock.Lock()
+	defer store.dbsLock.Unlock()
+
+	store.deleteTable(context.Background(), bucket)
+
+	if store.dbs == nil {
+		return
+	}
+	delete(store.dbs, bucket)
+}
+
 const (
 	DEFAULT_TABLE = "filemeta"
 )

weed/filer/filechunk_manifest.go

@@ -97,20 +97,20 @@ func fetchChunk(lookupFileIdFn wdclient.LookupFileIdFunctionType, fileId string,
 func retriedFetchChunkData(urlStrings []string, cipherKey []byte, isGzipped bool, isFullChunk bool, offset int64, size int) ([]byte, error) {
 
 	var err error
-	var buffer bytes.Buffer
 	var shouldRetry bool
+	receivedData := make([]byte, 0, size)
 
 	for waitTime := time.Second; waitTime < util.RetryWaitTime; waitTime += waitTime / 2 {
 		for _, urlString := range urlStrings {
-			shouldRetry, err = util.FastReadUrlAsStream(urlString+"?readDeleted=true", cipherKey, isGzipped, isFullChunk, offset, size, func(data []byte) {
-				buffer.Write(data)
+			receivedData = receivedData[:0]
+			shouldRetry, err = util.ReadUrlAsStream(urlString+"?readDeleted=true", cipherKey, isGzipped, isFullChunk, offset, size, func(data []byte) {
+				receivedData = append(receivedData, data...)
 			})
 			if !shouldRetry {
 				break
 			}
 			if err != nil {
 				glog.V(0).Infof("read %s failed, err: %v", urlString, err)
-				buffer.Reset()
 			} else {
 				break
 			}
@@ -123,7 +123,8 @@ func retriedFetchChunkData(urlStrings []string, cipherKey []byte, isGzipped bool
 		}
 	}
 
-	return buffer.Bytes(), err
+	return receivedData, err
+
 }
 
 func MaybeManifestize(saveFunc SaveDataAsChunkFunctionType, inputChunks []*filer_pb.FileChunk) (chunks []*filer_pb.FileChunk, err error) {

weed/filer/filer.go

@@ -241,12 +241,12 @@ func (f *Filer) UpdateEntry(ctx context.Context, oldEntry, entry *Entry) (err er
 	if oldEntry != nil {
 		entry.Attr.Crtime = oldEntry.Attr.Crtime
 		if oldEntry.IsDirectory() && !entry.IsDirectory() {
-			glog.Errorf("existing %s is a directory", entry.FullPath)
-			return fmt.Errorf("existing %s is a directory", entry.FullPath)
+			glog.Errorf("existing %s is a directory", oldEntry.FullPath)
+			return fmt.Errorf("existing %s is a directory", oldEntry.FullPath)
 		}
 		if !oldEntry.IsDirectory() && entry.IsDirectory() {
-			glog.Errorf("existing %s is a file", entry.FullPath)
-			return fmt.Errorf("existing %s is a file", entry.FullPath)
+			glog.Errorf("existing %s is a file", oldEntry.FullPath)
+			return fmt.Errorf("existing %s is a file", oldEntry.FullPath)
 		}
 	}
 	return f.Store.UpdateEntry(ctx, entry)

weed/filer/filer_delete_entry.go

@@ -11,6 +11,10 @@ import (
 
 type HardLinkId []byte
 
+const (
+	MsgFailDelNonEmptyFolder = "fail to delete non-empty folder"
+)
+
 func (f *Filer) DeleteEntryMetaAndData(ctx context.Context, p util.FullPath, isRecursive, ignoreRecursiveError, shouldDeleteChunks, isFromOtherCluster bool, signatures []int32) (err error) {
 	if p == "/" {
 		return nil
@@ -77,7 +81,7 @@ func (f *Filer) doBatchDeleteFolderMetaAndData(ctx context.Context, entry *Entry
 			if lastFileName == "" && !isRecursive && len(entries) > 0 {
 				// only for first iteration in the loop
 				glog.Errorf("deleting a folder %s has children: %+v ...", entry.FullPath, entries[0].Name())
-				return nil, nil, fmt.Errorf("fail to delete non-empty folder: %s", entry.FullPath)
+				return nil, nil, fmt.Errorf("%s: %s", MsgFailDelNonEmptyFolder, entry.FullPath)
 			}
 
 			for _, sub := range entries {

weed/filer/filer_on_meta_event.go

@@ -11,6 +11,28 @@ import (
 
 // onMetadataChangeEvent is triggered after filer processed change events from local or remote filers
 func (f *Filer) onMetadataChangeEvent(event *filer_pb.SubscribeMetadataResponse) {
+	f.maybeReloadFilerConfiguration(event)
+	f.onBucketEvents(event)
+}
+
+func (f *Filer) onBucketEvents(event *filer_pb.SubscribeMetadataResponse) {
+	message := event.EventNotification
+	for _, sig := range message.Signatures {
+		if sig == f.Signature {
+			return
+		}
+	}
+	if f.DirBucketsPath == event.Directory {
+		if message.OldEntry == nil && message.NewEntry != nil {
+			f.Store.OnBucketCreation(message.NewEntry.Name)
+		}
+		if message.OldEntry != nil && message.NewEntry == nil {
+			f.Store.OnBucketDeletion(message.OldEntry.Name)
+		}
+	}
+}
+
+func (f *Filer) maybeReloadFilerConfiguration(event *filer_pb.SubscribeMetadataResponse) {
 	if DirectoryEtcSeaweedFS != event.Directory {
 		if DirectoryEtcSeaweedFS != event.EventNotification.NewParentPath {
 			return
@@ -26,12 +48,11 @@ func (f *Filer) onMetadataChangeEvent(event *filer_pb.SubscribeMetadataResponse)
 	if entry.Name == FilerConfName {
 		f.reloadFilerConfiguration(entry)
 	}
-
 }
 
 func (f *Filer) readEntry(chunks []*filer_pb.FileChunk) ([]byte, error) {
 	var buf bytes.Buffer
-	err := StreamContent(f.MasterClient, &buf, chunks, 0, math.MaxInt64)
+	err := StreamContent(f.MasterClient, &buf, chunks, 0, math.MaxInt64, false)
 	if err != nil {
 		return nil, err
 	}

weed/filer/filerstore.go

@@ -39,3 +39,8 @@ type FilerStore interface {
 
 	Shutdown()
 }
+
+type BucketAware interface {
+	OnBucketCreation(bucket string)
+	OnBucketDeletion(bucket string)
+}

weed/filer/filerstore_wrapper.go

@@ -21,6 +21,8 @@ type VirtualFilerStore interface {
 	DeleteHardLink(ctx context.Context, hardLinkId HardLinkId) error
 	DeleteOneEntry(ctx context.Context, entry *Entry) error
 	AddPathSpecificStore(path string, storeId string, store FilerStore)
+	OnBucketCreation(bucket string)
+	OnBucketDeletion(bucket string)
 }
 
 type FilerStoreWrapper struct {
@@ -40,6 +42,27 @@ func NewFilerStoreWrapper(store FilerStore) *FilerStoreWrapper {
 	}
 }
 
+func (fsw *FilerStoreWrapper) OnBucketCreation(bucket string) {
+	for _, store := range fsw.storeIdToStore {
+		if ba, ok := store.(BucketAware); ok {
+			ba.OnBucketCreation(bucket)
+		}
+	}
+	if ba, ok := fsw.defaultStore.(BucketAware); ok {
+		ba.OnBucketCreation(bucket)
+	}
+}
+func (fsw *FilerStoreWrapper) OnBucketDeletion(bucket string) {
+	for _, store := range fsw.storeIdToStore {
+		if ba, ok := store.(BucketAware); ok {
+			ba.OnBucketDeletion(bucket)
+		}
+	}
+	if ba, ok := fsw.defaultStore.(BucketAware); ok {
+		ba.OnBucketDeletion(bucket)
+	}
+}
+
 func (fsw *FilerStoreWrapper) AddPathSpecificStore(path string, storeId string, store FilerStore) {
 	fsw.storeIdToStore[storeId] = NewFilerStorePathTranlator(path, store)
 	err := fsw.pathToStore.Put([]byte(path), storeId)

weed/filer/leveldb/leveldb_store.go

@@ -8,6 +8,7 @@ import (
 	leveldb_errors "github.com/syndtr/goleveldb/leveldb/errors"
 	"github.com/syndtr/goleveldb/leveldb/opt"
 	leveldb_util "github.com/syndtr/goleveldb/leveldb/util"
+	"os"
 
 	"github.com/chrislusf/seaweedfs/weed/filer"
 	"github.com/chrislusf/seaweedfs/weed/glog"
@@ -38,6 +39,7 @@ func (store *LevelDBStore) Initialize(configuration weed_util.Configuration, pre
 
 func (store *LevelDBStore) initialize(dir string) (err error) {
 	glog.Infof("filer store dir: %s", dir)
+	os.MkdirAll(dir, 0755)
 	if err := weed_util.TestFolderWritable(dir); err != nil {
 		return fmt.Errorf("Check Level Folder %s Writable: %s", dir, err)
 	}

weed/filer/leveldb2/leveldb2_store.go

@@ -38,6 +38,7 @@ func (store *LevelDB2Store) Initialize(configuration weed_util.Configuration, pr
 
 func (store *LevelDB2Store) initialize(dir string, dbCount int) (err error) {
 	glog.Infof("filer store leveldb2 dir: %s", dir)
+	os.MkdirAll(dir, 0755)
 	if err := weed_util.TestFolderWritable(dir); err != nil {
 		return fmt.Errorf("Check Level Folder %s Writable: %s", dir, err)
 	}

weed/filer/leveldb3/leveldb3_store.go

@@ -45,6 +45,7 @@ func (store *LevelDB3Store) Initialize(configuration weed_util.Configuration, pr
 
 func (store *LevelDB3Store) initialize(dir string) (err error) {
 	glog.Infof("filer store leveldb3 dir: %s", dir)
+	os.MkdirAll(dir, 0755)
 	if err := weed_util.TestFolderWritable(dir); err != nil {
 		return fmt.Errorf("Check Level Folder %s Writable: %s", dir, err)
 	}

weed/filer/meta_aggregator.go

@@ -69,6 +69,7 @@ func (ma *MetaAggregator) subscribeToOneFiler(f *Filer, self string, peer string
 		peerSignature, err = ma.readFilerStoreSignature(peer)
 	}
 
+	// when filer store is not shared by multiple filers
 	if peerSignature != f.Signature {
 		if prevTsNs, err := ma.readOffset(f, peer, peerSignature); err == nil {
 			lastTsNs = prevTsNs

weed/filer/mysql/mysql_sql_gen.go

@@ -2,6 +2,7 @@ package mysql
 
 import (
 	"fmt"
+
 	"github.com/chrislusf/seaweedfs/weed/filer/abstract_sql"
 	_ "github.com/go-sql-driver/mysql"
 )
@@ -9,44 +10,49 @@ import (
 type SqlGenMysql struct {
 	CreateTableSqlTemplate string
 	DropTableSqlTemplate   string
+	UpsertQueryTemplate    string
 }
 
 var (
 	_ = abstract_sql.SqlGenerator(&SqlGenMysql{})
 )
 
-func (gen *SqlGenMysql) GetSqlInsert(bucket string) string {
-	return fmt.Sprintf("INSERT INTO `%s` (dirhash,name,directory,meta) VALUES(?,?,?,?)", bucket)
+func (gen *SqlGenMysql) GetSqlInsert(tableName string) string {
+	if gen.UpsertQueryTemplate != "" {
+		return fmt.Sprintf(gen.UpsertQueryTemplate, tableName)
+	} else {
+		return fmt.Sprintf("INSERT INTO `%s` (dirhash,name,directory,meta) VALUES(?,?,?,?)", tableName)
+	}
 }
 
-func (gen *SqlGenMysql) GetSqlUpdate(bucket string) string {
-	return fmt.Sprintf("UPDATE `%s` SET meta=? WHERE dirhash=? AND name=? AND directory=?", bucket)
+func (gen *SqlGenMysql) GetSqlUpdate(tableName string) string {
+	return fmt.Sprintf("UPDATE `%s` SET meta=? WHERE dirhash=? AND name=? AND directory=?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlFind(bucket string) string {
-	return fmt.Sprintf("SELECT meta FROM `%s` WHERE dirhash=? AND name=? AND directory=?", bucket)
+func (gen *SqlGenMysql) GetSqlFind(tableName string) string {
+	return fmt.Sprintf("SELECT meta FROM `%s` WHERE dirhash=? AND name=? AND directory=?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlDelete(bucket string) string {
-	return fmt.Sprintf("DELETE FROM `%s` WHERE dirhash=? AND name=? AND directory=?", bucket)
+func (gen *SqlGenMysql) GetSqlDelete(tableName string) string {
+	return fmt.Sprintf("DELETE FROM `%s` WHERE dirhash=? AND name=? AND directory=?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlDeleteFolderChildren(bucket string) string {
-	return fmt.Sprintf("DELETE FROM `%s` WHERE dirhash=? AND directory=?", bucket)
+func (gen *SqlGenMysql) GetSqlDeleteFolderChildren(tableName string) string {
+	return fmt.Sprintf("DELETE FROM `%s` WHERE dirhash=? AND directory=?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlListExclusive(bucket string) string {
-	return fmt.Sprintf("SELECT NAME, meta FROM `%s` WHERE dirhash=? AND name>? AND directory=? AND name like ? ORDER BY NAME ASC LIMIT ?", bucket)
+func (gen *SqlGenMysql) GetSqlListExclusive(tableName string) string {
+	return fmt.Sprintf("SELECT NAME, meta FROM `%s` WHERE dirhash=? AND name>? AND directory=? AND name like ? ORDER BY NAME ASC LIMIT ?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlListInclusive(bucket string) string {
-	return fmt.Sprintf("SELECT NAME, meta FROM `%s` WHERE dirhash=? AND name>=? AND directory=? AND name like ? ORDER BY NAME ASC LIMIT ?", bucket)
+func (gen *SqlGenMysql) GetSqlListInclusive(tableName string) string {
+	return fmt.Sprintf("SELECT NAME, meta FROM `%s` WHERE dirhash=? AND name>=? AND directory=? AND name like ? ORDER BY NAME ASC LIMIT ?", tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlCreateTable(bucket string) string {
-	return fmt.Sprintf(gen.CreateTableSqlTemplate, bucket)
+func (gen *SqlGenMysql) GetSqlCreateTable(tableName string) string {
+	return fmt.Sprintf(gen.CreateTableSqlTemplate, tableName)
 }
 
-func (gen *SqlGenMysql) GetSqlDropTable(bucket string) string {
-	return fmt.Sprintf(gen.DropTableSqlTemplate, bucket)
+func (gen *SqlGenMysql) GetSqlDropTable(tableName string) string {
+	return fmt.Sprintf(gen.DropTableSqlTemplate, tableName)
 }

weed/filer/mysql/mysql_store.go

@@ -3,9 +3,10 @@ package mysql
 import (
 	"database/sql"
 	"fmt"
-	"github.com/chrislusf/seaweedfs/weed/filer"
 	"time"
 
+	"github.com/chrislusf/seaweedfs/weed/filer"
+
 	"github.com/chrislusf/seaweedfs/weed/filer/abstract_sql"
 	"github.com/chrislusf/seaweedfs/weed/util"
 	_ "github.com/go-sql-driver/mysql"
@@ -29,6 +30,8 @@ func (store *MysqlStore) GetName() string {
 
 func (store *MysqlStore) Initialize(configuration util.Configuration, prefix string) (err error) {
 	return store.initialize(
+		configuration.GetString(prefix+"upsertQuery"),
+		configuration.GetBool(prefix+"enableUpsert"),
 		configuration.GetString(prefix+"username"),
 		configuration.GetString(prefix+"password"),
 		configuration.GetString(prefix+"hostname"),
@@ -41,13 +44,17 @@ func (store *MysqlStore) Initialize(configuration util.Configuration, prefix str
 	)
 }
 
-func (store *MysqlStore) initialize(user, password, hostname string, port int, database string, maxIdle, maxOpen,
+func (store *MysqlStore) initialize(upsertQuery string, enableUpsert bool, user, password, hostname string, port int, database string, maxIdle, maxOpen,
 	maxLifetimeSeconds int, interpolateParams bool) (err error) {
 
 	store.SupportBucketTable = false
+	if !enableUpsert {
+		upsertQuery = ""
+	}
 	store.SqlGenerator = &SqlGenMysql{
 		CreateTableSqlTemplate: "",
 		DropTableSqlTemplate:   "drop table `%s`",
+		UpsertQueryTemplate:    upsertQuery,
 	}
 
 	sqlUrl := fmt.Sprintf(CONNECTION_URL_PATTERN, user, password, hostname, port, database)

weed/filer/mysql2/mysql2_store.go

@@ -32,6 +32,8 @@ func (store *MysqlStore2) GetName() string {
 func (store *MysqlStore2) Initialize(configuration util.Configuration, prefix string) (err error) {
 	return store.initialize(
 		configuration.GetString(prefix+"createTable"),
+		configuration.GetString(prefix+"upsertQuery"),
+		configuration.GetBool(prefix+"enableUpsert"),
 		configuration.GetString(prefix+"username"),
 		configuration.GetString(prefix+"password"),
 		configuration.GetString(prefix+"hostname"),
@@ -44,13 +46,17 @@ func (store *MysqlStore2) Initialize(configuration util.Configuration, prefix st
 	)
 }
 
-func (store *MysqlStore2) initialize(createTable, user, password, hostname string, port int, database string, maxIdle, maxOpen,
+func (store *MysqlStore2) initialize(createTable, upsertQuery string, enableUpsert bool, user, password, hostname string, port int, database string, maxIdle, maxOpen,
 	maxLifetimeSeconds int, interpolateParams bool) (err error) {
 
 	store.SupportBucketTable = true
+	if !enableUpsert {
+		upsertQuery = ""
+	}
 	store.SqlGenerator = &mysql.SqlGenMysql{
 		CreateTableSqlTemplate: createTable,
 		DropTableSqlTemplate:   "drop table `%s`",
+		UpsertQueryTemplate:    upsertQuery,
 	}
 
 	sqlUrl := fmt.Sprintf(CONNECTION_URL_PATTERN, user, password, hostname, port, database)

weed/filer/postgres/postgres_sql_gen.go

@@ -10,44 +10,49 @@ import (
 type SqlGenPostgres struct {
 	CreateTableSqlTemplate string
 	DropTableSqlTemplate   string
+	UpsertQueryTemplate    string
 }
 
 var (
 	_ = abstract_sql.SqlGenerator(&SqlGenPostgres{})
 )
 
-func (gen *SqlGenPostgres) GetSqlInsert(bucket string) string {
-	return fmt.Sprintf(`INSERT INTO "%s" (dirhash,name,directory,meta) VALUES($1,$2,$3,$4)`, bucket)
+func (gen *SqlGenPostgres) GetSqlInsert(tableName string) string {
+	if gen.UpsertQueryTemplate != "" {
+		return fmt.Sprintf(gen.UpsertQueryTemplate, tableName)
+	} else {
+		return fmt.Sprintf(`INSERT INTO "%s" (dirhash,name,directory,meta) VALUES($1,$2,$3,$4)`, tableName)
+	}
 }
 
-func (gen *SqlGenPostgres) GetSqlUpdate(bucket string) string {
-	return fmt.Sprintf(`UPDATE "%s" SET meta=$1 WHERE dirhash=$2 AND name=$3 AND directory=$4`, bucket)
+func (gen *SqlGenPostgres) GetSqlUpdate(tableName string) string {
+	return fmt.Sprintf(`UPDATE "%s" SET meta=$1 WHERE dirhash=$2 AND name=$3 AND directory=$4`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlFind(bucket string) string {
-	return fmt.Sprintf(`SELECT meta FROM "%s" WHERE dirhash=$1 AND name=$2 AND directory=$3`, bucket)
+func (gen *SqlGenPostgres) GetSqlFind(tableName string) string {
+	return fmt.Sprintf(`SELECT meta FROM "%s" WHERE dirhash=$1 AND name=$2 AND directory=$3`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlDelete(bucket string) string {
-	return fmt.Sprintf(`DELETE FROM "%s" WHERE dirhash=$1 AND name=$2 AND directory=$3`, bucket)
+func (gen *SqlGenPostgres) GetSqlDelete(tableName string) string {
+	return fmt.Sprintf(`DELETE FROM "%s" WHERE dirhash=$1 AND name=$2 AND directory=$3`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlDeleteFolderChildren(bucket string) string {
-	return fmt.Sprintf(`DELETE FROM "%s" WHERE dirhash=$1 AND directory=$2`, bucket)
+func (gen *SqlGenPostgres) GetSqlDeleteFolderChildren(tableName string) string {
+	return fmt.Sprintf(`DELETE FROM "%s" WHERE dirhash=$1 AND directory=$2`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlListExclusive(bucket string) string {
-	return fmt.Sprintf(`SELECT NAME, meta FROM "%s" WHERE dirhash=$1 AND name>$2 AND directory=$3 AND name like $4 ORDER BY NAME ASC LIMIT $5`, bucket)
+func (gen *SqlGenPostgres) GetSqlListExclusive(tableName string) string {
+	return fmt.Sprintf(`SELECT NAME, meta FROM "%s" WHERE dirhash=$1 AND name>$2 AND directory=$3 AND name like $4 ORDER BY NAME ASC LIMIT $5`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlListInclusive(bucket string) string {
-	return fmt.Sprintf(`SELECT NAME, meta FROM "%s" WHERE dirhash=$1 AND name>=$2 AND directory=$3 AND name like $4 ORDER BY NAME ASC LIMIT $5`, bucket)
+func (gen *SqlGenPostgres) GetSqlListInclusive(tableName string) string {
+	return fmt.Sprintf(`SELECT NAME, meta FROM "%s" WHERE dirhash=$1 AND name>=$2 AND directory=$3 AND name like $4 ORDER BY NAME ASC LIMIT $5`, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlCreateTable(bucket string) string {
-	return fmt.Sprintf(gen.CreateTableSqlTemplate, bucket)
+func (gen *SqlGenPostgres) GetSqlCreateTable(tableName string) string {
+	return fmt.Sprintf(gen.CreateTableSqlTemplate, tableName)
 }
 
-func (gen *SqlGenPostgres) GetSqlDropTable(bucket string) string {
-	return fmt.Sprintf(gen.DropTableSqlTemplate, bucket)
+func (gen *SqlGenPostgres) GetSqlDropTable(tableName string) string {
+	return fmt.Sprintf(gen.DropTableSqlTemplate, tableName)
 }

weed/filer/postgres/postgres_store.go

@@ -29,6 +29,8 @@ func (store *PostgresStore) GetName() string {
 
 func (store *PostgresStore) Initialize(configuration util.Configuration, prefix string) (err error) {
 	return store.initialize(
+		configuration.GetString(prefix+"upsertQuery"),
+		configuration.GetBool(prefix+"enableUpsert"),
 		configuration.GetString(prefix+"username"),
 		configuration.GetString(prefix+"password"),
 		configuration.GetString(prefix+"hostname"),
@@ -42,12 +44,16 @@ func (store *PostgresStore) Initialize(configuration util.Configuration, prefix
 	)
 }
 
-func (store *PostgresStore) initialize(user, password, hostname string, port int, database, schema, sslmode string, maxIdle, maxOpen, maxLifetimeSeconds int) (err error) {
+func (store *PostgresStore) initialize(upsertQuery string, enableUpsert bool, user, password, hostname string, port int, database, schema, sslmode string, maxIdle, maxOpen, maxLifetimeSeconds int) (err error) {
 
 	store.SupportBucketTable = false
+	if !enableUpsert {
+		upsertQuery = ""
+	}
 	store.SqlGenerator = &SqlGenPostgres{
 		CreateTableSqlTemplate: "",
 		DropTableSqlTemplate:   `drop table "%s"`,
+		UpsertQueryTemplate:    upsertQuery,
 	}
 
 	sqlUrl := fmt.Sprintf(CONNECTION_URL_PATTERN, hostname, port, sslmode)

weed/filer/postgres2/postgres2_store.go

@@ -32,6 +32,8 @@ func (store *PostgresStore2) GetName() string {
 func (store *PostgresStore2) Initialize(configuration util.Configuration, prefix string) (err error) {
 	return store.initialize(
 		configuration.GetString(prefix+"createTable"),
+		configuration.GetString(prefix+"upsertQuery"),
+		configuration.GetBool(prefix+"enableUpsert"),
 		configuration.GetString(prefix+"username"),
 		configuration.GetString(prefix+"password"),
 		configuration.GetString(prefix+"hostname"),
@@ -45,12 +47,16 @@ func (store *PostgresStore2) Initialize(configuration util.Configuration, prefix
 	)
 }
 
-func (store *PostgresStore2) initialize(createTable, user, password, hostname string, port int, database, schema, sslmode string, maxIdle, maxOpen, maxLifetimeSeconds int) (err error) {
+func (store *PostgresStore2) initialize(createTable, upsertQuery string, enableUpsert bool, user, password, hostname string, port int, database, schema, sslmode string, maxIdle, maxOpen, maxLifetimeSeconds int) (err error) {
 
 	store.SupportBucketTable = true
+	if !enableUpsert {
+		upsertQuery = ""
+	}
 	store.SqlGenerator = &postgres.SqlGenPostgres{
 		CreateTableSqlTemplate: createTable,
 		DropTableSqlTemplate:   `drop table "%s"`,
+		UpsertQueryTemplate:    upsertQuery,
 	}
 
 	sqlUrl := fmt.Sprintf(CONNECTION_URL_PATTERN, hostname, port, sslmode)

weed/filer/read_write.go

@@ -27,7 +27,7 @@ func ReadEntry(masterClient *wdclient.MasterClient, filerClient filer_pb.Seaweed
 		return err
 	}
 
-	return StreamContent(masterClient, byteBuffer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64)
+	return StreamContent(masterClient, byteBuffer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64, false)
 
 }
 
@@ -35,7 +35,7 @@ func ReadContent(filerAddress string, dir, name string) ([]byte, error) {
 
 	target := fmt.Sprintf("http://%s%s/%s", filerAddress, dir, name)
 
-	data, _, err := util.FastGet(target)
+	data, _, err := util.Get(target)
 
 	return data, err
 }

weed/filer/rocksdb/rocksdb_store.go

@@ -8,6 +8,7 @@ import (
 	"crypto/md5"
 	"fmt"
 	"io"
+	"os"
 
 	"github.com/tecbot/gorocksdb"
 
@@ -56,6 +57,7 @@ func (store *RocksDBStore) Initialize(configuration weed_util.Configuration, pre
 
 func (store *RocksDBStore) initialize(dir string) (err error) {
 	glog.Infof("filer store rocksdb dir: %s", dir)
+	os.MkdirAll(dir, 0755)
 	if err := weed_util.TestFolderWritable(dir); err != nil {
 		return fmt.Errorf("Check Level Folder %s Writable: %s", dir, err)
 	}

weed/filer/stream.go

@@ -3,6 +3,7 @@ package filer
 import (
 	"bytes"
 	"fmt"
+	"golang.org/x/sync/errgroup"
 	"io"
 	"math"
 	"strings"
@@ -13,9 +14,9 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/wdclient"
 )
 
-func StreamContent(masterClient wdclient.HasLookupFileIdFunction, w io.Writer, chunks []*filer_pb.FileChunk, offset int64, size int64) error {
+func StreamContent(masterClient wdclient.HasLookupFileIdFunction, w io.Writer, chunks []*filer_pb.FileChunk, offset int64, size int64, isCheck bool) error {
 
-	// fmt.Printf("start to stream content for chunks: %+v\n", chunks)
+	glog.V(9).Infof("start to stream content for chunks: %+v\n", chunks)
 	chunkViews := ViewFromChunks(masterClient.GetLookupFileIdFunction(), chunks, offset, size)
 
 	fileId2Url := make(map[string][]string)
@@ -26,19 +27,33 @@ func StreamContent(masterClient wdclient.HasLookupFileIdFunction, w io.Writer, c
 		if err != nil {
 			glog.V(1).Infof("operation LookupFileId %s failed, err: %v", chunkView.FileId, err)
 			return err
+		} else if len(urlStrings) == 0 {
+			glog.Errorf("operation LookupFileId %s failed, err: urls not found", chunkView.FileId)
+			return fmt.Errorf("operation LookupFileId %s failed, err: urls not found", chunkView.FileId)
 		}
 		fileId2Url[chunkView.FileId] = urlStrings
 	}
 
+	if isCheck {
+		// Pre-check all chunkViews urls
+		gErr := new(errgroup.Group)
+		CheckAllChunkViews(chunkViews, &fileId2Url, gErr)
+		if err := gErr.Wait(); err != nil {
+			glog.Errorf("check all chunks: %v", err)
+			return fmt.Errorf("check all chunks: %v", err)
+		}
+		return nil
+	}
+
 	for _, chunkView := range chunkViews {
 
 		urlStrings := fileId2Url[chunkView.FileId]
-
 		data, err := retriedFetchChunkData(urlStrings, chunkView.CipherKey, chunkView.IsGzipped, chunkView.IsFullChunk(), chunkView.Offset, int(chunkView.Size))
 		if err != nil {
 			glog.Errorf("read chunk: %v", err)
 			return fmt.Errorf("read chunk: %v", err)
 		}
+
 		_, err = w.Write(data)
 		if err != nil {
 			glog.Errorf("write chunk: %v", err)
@@ -50,6 +65,17 @@ func StreamContent(masterClient wdclient.HasLookupFileIdFunction, w io.Writer, c
 
 }
 
+func CheckAllChunkViews(chunkViews []*ChunkView, fileId2Url *map[string][]string, gErr *errgroup.Group) {
+	for _, chunkView := range chunkViews {
+		urlStrings := (*fileId2Url)[chunkView.FileId]
+		glog.V(9).Infof("Check chunk: %+v\n url: %v", chunkView, urlStrings)
+		gErr.Go(func() error {
+			_, err := retriedFetchChunkData(urlStrings, chunkView.CipherKey, chunkView.IsGzipped, chunkView.IsFullChunk(), chunkView.Offset, int(chunkView.Size))
+			return err
+		})
+	}
+}
+
 // ----------------  ReadAllReader ----------------------------------
 
 func ReadAll(masterClient *wdclient.MasterClient, chunks []*filer_pb.FileChunk) ([]byte, error) {
@@ -181,7 +207,7 @@ func (c *ChunkStreamReader) fetchChunkToBuffer(chunkView *ChunkView) error {
 	var buffer bytes.Buffer
 	var shouldRetry bool
 	for _, urlString := range urlStrings {
-		shouldRetry, err = util.FastReadUrlAsStream(urlString+"?readDeleted=true", chunkView.CipherKey, chunkView.IsGzipped, chunkView.IsFullChunk(), chunkView.Offset, int(chunkView.Size), func(data []byte) {
+		shouldRetry, err = util.ReadUrlAsStream(urlString, chunkView.CipherKey, chunkView.IsGzipped, chunkView.IsFullChunk(), chunkView.Offset, int(chunkView.Size), func(data []byte) {
 			buffer.Write(data)
 		})
 		if !shouldRetry {

weed/filesys/dir.go

@@ -128,6 +128,10 @@ func (dir *Dir) newDirectory(fullpath util.FullPath, entry *filer_pb.Entry) fs.N
 func (dir *Dir) Create(ctx context.Context, req *fuse.CreateRequest,
 	resp *fuse.CreateResponse) (fs.Node, fs.Handle, error) {
 
+	if dir.wfs.option.ReadOnly {
+		return nil, nil, fuse.EPERM
+	}
+
 	request, err := dir.doCreateEntry(req.Name, req.Mode, req.Uid, req.Gid, req.Flags&fuse.OpenExclusive != 0)
 
 	if err != nil {
@@ -148,6 +152,10 @@ func (dir *Dir) Create(ctx context.Context, req *fuse.CreateRequest,
 
 func (dir *Dir) Mknod(ctx context.Context, req *fuse.MknodRequest) (fs.Node, error) {
 
+	if dir.wfs.option.ReadOnly {
+		return nil, fuse.EPERM
+	}
+
 	request, err := dir.doCreateEntry(req.Name, req.Mode, req.Uid, req.Gid, false)
 
 	if err != nil {
@@ -202,6 +210,10 @@ func (dir *Dir) doCreateEntry(name string, mode os.FileMode, uid, gid uint32, ex
 
 func (dir *Dir) Mkdir(ctx context.Context, req *fuse.MkdirRequest) (fs.Node, error) {
 
+	if dir.wfs.option.ReadOnly {
+		return nil, fuse.EPERM
+	}
+
 	glog.V(4).Infof("mkdir %s: %s", dir.FullPath(), req.Name)
 
 	newEntry := &filer_pb.Entry{
@@ -251,10 +263,10 @@ func (dir *Dir) Mkdir(ctx context.Context, req *fuse.MkdirRequest) (fs.Node, err
 
 func (dir *Dir) Lookup(ctx context.Context, req *fuse.LookupRequest, resp *fuse.LookupResponse) (node fs.Node, err error) {
 
-	glog.V(4).Infof("dir Lookup %s: %s by %s", dir.FullPath(), req.Name, req.Header.String())
-
-	fullFilePath := util.NewFullPath(dir.FullPath(), req.Name)
 	dirPath := util.FullPath(dir.FullPath())
+	glog.V(4).Infof("dir Lookup %s: %s by %s", dirPath, req.Name, req.Header.String())
+
+	fullFilePath := dirPath.Child(req.Name)
 	visitErr := meta_cache.EnsureVisited(dir.wfs.metaCache, dir.wfs, dirPath)
 	if visitErr != nil {
 		glog.Errorf("dir Lookup %s: %v", dirPath, visitErr)
@@ -305,7 +317,8 @@ func (dir *Dir) Lookup(ctx context.Context, req *fuse.LookupRequest, resp *fuse.
 
 func (dir *Dir) ReadDirAll(ctx context.Context) (ret []fuse.Dirent, err error) {
 
-	glog.V(4).Infof("dir ReadDirAll %s", dir.FullPath())
+	dirPath := util.FullPath(dir.FullPath())
+	glog.V(4).Infof("dir ReadDirAll %s", dirPath)
 
 	processEachEntryFn := func(entry *filer_pb.Entry, isLast bool) error {
 		if entry.IsDirectory {
@@ -318,12 +331,11 @@ func (dir *Dir) ReadDirAll(ctx context.Context) (ret []fuse.Dirent, err error) {
 		return nil
 	}
 
-	dirPath := util.FullPath(dir.FullPath())
 	if err = meta_cache.EnsureVisited(dir.wfs.metaCache, dir.wfs, dirPath); err != nil {
 		glog.Errorf("dir ReadDirAll %s: %v", dirPath, err)
 		return nil, fuse.EIO
 	}
-	listErr := dir.wfs.metaCache.ListDirectoryEntries(context.Background(), util.FullPath(dir.FullPath()), "", false, int64(math.MaxInt32), func(entry *filer.Entry) bool {
+	listErr := dir.wfs.metaCache.ListDirectoryEntries(context.Background(), dirPath, "", false, int64(math.MaxInt32), func(entry *filer.Entry) bool {
 		processEachEntryFn(entry.ToProtoEntry(), false)
 		return true
 	})
@@ -356,6 +368,11 @@ func findFileType(mode uint16) fuse.DirentType {
 
 func (dir *Dir) Remove(ctx context.Context, req *fuse.RemoveRequest) error {
 
+	if dir.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
+
 	if !req.Dir {
 		return dir.removeOneFile(req)
 	}
@@ -389,12 +406,12 @@ func (dir *Dir) removeOneFile(req *fuse.RemoveRequest) error {
 
 	// clear entry inside the file
 	fsNode := dir.wfs.fsNodeCache.GetFsNode(filePath)
+	dir.wfs.fsNodeCache.DeleteFsNode(filePath)
 	if fsNode != nil {
 		if file, ok := fsNode.(*File); ok {
 			file.clearEntry()
 		}
 	}
-	dir.wfs.fsNodeCache.DeleteFsNode(filePath)
 
 	// remove current file handle if any
 	dir.wfs.handlesLock.Lock()
@@ -429,6 +446,10 @@ func (dir *Dir) removeFolder(req *fuse.RemoveRequest) error {
 
 func (dir *Dir) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *fuse.SetattrResponse) error {
 
+	if dir.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("%v dir setattr %+v", dir.FullPath(), req)
 
 	if err := dir.maybeLoadEntry(); err != nil {
@@ -457,6 +478,10 @@ func (dir *Dir) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *fus
 
 func (dir *Dir) Setxattr(ctx context.Context, req *fuse.SetxattrRequest) error {
 
+	if dir.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("dir Setxattr %s: %s", dir.FullPath(), req.Name)
 
 	if err := dir.maybeLoadEntry(); err != nil {
@@ -473,6 +498,10 @@ func (dir *Dir) Setxattr(ctx context.Context, req *fuse.SetxattrRequest) error {
 
 func (dir *Dir) Removexattr(ctx context.Context, req *fuse.RemovexattrRequest) error {
 
+	if dir.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("dir Removexattr %s: %s", dir.FullPath(), req.Name)
 
 	if err := dir.maybeLoadEntry(); err != nil {

weed/filesys/dir_link.go

@@ -24,6 +24,10 @@ const (
 
 func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (fs.Node, error) {
 
+	if dir.wfs.option.ReadOnly {
+		return nil, fuse.EPERM
+	}
+
 	oldFile, ok := old.(*File)
 	if !ok {
 		glog.Errorf("old node is not a file: %+v", old)
@@ -35,15 +39,20 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 		return nil, err
 	}
 
-	// update old file to hardlink mode
-	if len(oldFile.entry.HardLinkId) == 0 {
-		oldFile.entry.HardLinkId = append(util.RandomBytes(16), HARD_LINK_MARKER)
-		oldFile.entry.HardLinkCounter = 1
+	oldEntry := oldFile.getEntry()
+	if oldEntry == nil {
+		return nil, fuse.EIO
 	}
-	oldFile.entry.HardLinkCounter++
+
+	// update old file to hardlink mode
+	if len(oldEntry.HardLinkId) == 0 {
+		oldEntry.HardLinkId = append(util.RandomBytes(16), HARD_LINK_MARKER)
+		oldEntry.HardLinkCounter = 1
+	}
+	oldEntry.HardLinkCounter++
 	updateOldEntryRequest := &filer_pb.UpdateEntryRequest{
 		Directory:  oldFile.dir.FullPath(),
-		Entry:      oldFile.entry,
+		Entry:      oldEntry,
 		Signatures: []int32{dir.wfs.signature},
 	}
 
@@ -53,11 +62,11 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 		Entry: &filer_pb.Entry{
 			Name:            req.NewName,
 			IsDirectory:     false,
-			Attributes:      oldFile.entry.Attributes,
-			Chunks:          oldFile.entry.Chunks,
-			Extended:        oldFile.entry.Extended,
-			HardLinkId:      oldFile.entry.HardLinkId,
-			HardLinkCounter: oldFile.entry.HardLinkCounter,
+			Attributes:      oldEntry.Attributes,
+			Chunks:          oldEntry.Chunks,
+			Extended:        oldEntry.Extended,
+			HardLinkId:      oldEntry.HardLinkId,
+			HardLinkCounter: oldEntry.HardLinkCounter,
 		},
 		Signatures: []int32{dir.wfs.signature},
 	}
@@ -83,6 +92,10 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 		return nil
 	})
 
+	if err != nil {
+		return nil, fuse.EIO
+	}
+
 	// create new file node
 	newNode := dir.newFile(req.NewName, request.Entry)
 	newFile := newNode.(*File)
@@ -96,6 +109,10 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 
 func (dir *Dir) Symlink(ctx context.Context, req *fuse.SymlinkRequest) (fs.Node, error) {
 
+	if dir.wfs.option.ReadOnly {
+		return nil, fuse.EPERM
+	}
+
 	glog.V(4).Infof("Symlink: %v/%v to %v", dir.FullPath(), req.NewName, req.Target)
 
 	request := &filer_pb.CreateEntryRequest{

weed/filesys/dir_rename.go

@@ -13,6 +13,10 @@ import (
 
 func (dir *Dir) Rename(ctx context.Context, req *fuse.RenameRequest, newDirectory fs.Node) error {
 
+	if dir.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	newDir := newDirectory.(*Dir)
 
 	newPath := util.NewFullPath(newDir.FullPath(), req.NewName)

weed/filesys/dir_test.go

@@ -1,34 +0,0 @@
-package filesys
-
-import (
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-)
-
-func TestDirPath(t *testing.T) {
-
-	p := &Dir{name: "/some"}
-	p = &Dir{name: "path", parent: p}
-	p = &Dir{name: "to", parent: p}
-	p = &Dir{name: "a", parent: p}
-	p = &Dir{name: "file", parent: p}
-
-	assert.Equal(t, "/some/path/to/a/file", p.FullPath())
-
-	p = &Dir{name: "/some"}
-	assert.Equal(t, "/some", p.FullPath())
-
-	p = &Dir{name: "/"}
-	assert.Equal(t, "/", p.FullPath())
-
-	p = &Dir{name: "/"}
-	p = &Dir{name: "path", parent: p}
-	assert.Equal(t, "/path", p.FullPath())
-
-	p = &Dir{name: "/"}
-	p = &Dir{name: "path", parent: p}
-	p = &Dir{name: "to", parent: p}
-	assert.Equal(t, "/path/to", p.FullPath())
-
-}

weed/filesys/dirty_page.go

@@ -30,7 +30,7 @@ func newDirtyPages(file *File) *ContinuousDirtyPages {
 
 func (pages *ContinuousDirtyPages) AddPage(offset int64, data []byte) {
 
-	glog.V(4).Infof("%s AddPage [%d,%d) of %d bytes", pages.f.fullpath(), offset, offset+int64(len(data)), pages.f.entry.Attributes.FileSize)
+	glog.V(4).Infof("%s AddPage [%d,%d)", pages.f.fullpath(), offset, offset+int64(len(data)))
 
 	if len(data) > int(pages.f.wfs.option.ChunkSizeLimit) {
 		// this is more than what buffer can hold.
@@ -69,7 +69,12 @@ func (pages *ContinuousDirtyPages) saveExistingLargestPageToStorage() (hasSavedD
 		return false
 	}
 
-	fileSize := int64(pages.f.entry.Attributes.FileSize)
+	entry := pages.f.getEntry()
+	if entry == nil {
+		return false
+	}
+
+	fileSize := int64(entry.Attributes.FileSize)
 
 	chunkSize := min(maxList.Size(), fileSize-maxList.Offset())
 	if chunkSize == 0 {

weed/filesys/file.go

@@ -5,6 +5,7 @@ import (
 	"io"
 	"os"
 	"sort"
+	"sync"
 	"time"
 
 	"github.com/seaweedfs/fuse"
@@ -33,6 +34,7 @@ type File struct {
 	dir            *Dir
 	wfs            *WFS
 	entry          *filer_pb.Entry
+	entryLock      sync.RWMutex
 	entryViewCache []filer.VisibleInterval
 	isOpen         int
 	reader         io.ReaderAt
@@ -47,13 +49,17 @@ func (file *File) Attr(ctx context.Context, attr *fuse.Attr) (err error) {
 
 	glog.V(4).Infof("file Attr %s, open:%v existing:%v", file.fullpath(), file.isOpen, attr)
 
-	entry := file.entry
+	entry := file.getEntry()
 	if file.isOpen <= 0 || entry == nil {
 		if entry, err = file.maybeLoadEntry(ctx); err != nil {
 			return err
 		}
 	}
 
+	if entry == nil {
+		return fuse.ENOENT
+	}
+
 	// attr.Inode = file.fullpath().AsInode()
 	attr.Valid = time.Second
 	attr.Mode = os.FileMode(entry.Attributes.FileMode)
@@ -104,9 +110,13 @@ func (file *File) Open(ctx context.Context, req *fuse.OpenRequest, resp *fuse.Op
 
 func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *fuse.SetattrResponse) error {
 
+	if file.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("%v file setattr %+v", file.fullpath(), req)
 
-	_, err := file.maybeLoadEntry(ctx)
+	entry, err := file.maybeLoadEntry(ctx)
 	if err != nil {
 		return err
 	}
@@ -123,12 +133,12 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 
 	if req.Valid.Size() {
 
-		glog.V(4).Infof("%v file setattr set size=%v chunks=%d", file.fullpath(), req.Size, len(file.entry.Chunks))
-		if req.Size < filer.FileSize(file.entry) {
+		glog.V(4).Infof("%v file setattr set size=%v chunks=%d", file.fullpath(), req.Size, len(entry.Chunks))
+		if req.Size < filer.FileSize(entry) {
 			// fmt.Printf("truncate %v \n", fullPath)
 			var chunks []*filer_pb.FileChunk
 			var truncatedChunks []*filer_pb.FileChunk
-			for _, chunk := range file.entry.Chunks {
+			for _, chunk := range entry.Chunks {
 				int64Size := int64(chunk.Size)
 				if chunk.Offset+int64Size > int64(req.Size) {
 					// this chunk is truncated
@@ -143,36 +153,36 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 					}
 				}
 			}
-			file.entry.Chunks = chunks
+			entry.Chunks = chunks
 			file.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(file.wfs.LookupFn(), chunks)
-			file.reader = nil
+			file.setReader(nil)
 		}
-		file.entry.Attributes.FileSize = req.Size
+		entry.Attributes.FileSize = req.Size
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Mode() {
-		file.entry.Attributes.FileMode = uint32(req.Mode)
+		entry.Attributes.FileMode = uint32(req.Mode)
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Uid() {
-		file.entry.Attributes.Uid = req.Uid
+		entry.Attributes.Uid = req.Uid
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Gid() {
-		file.entry.Attributes.Gid = req.Gid
+		entry.Attributes.Gid = req.Gid
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Crtime() {
-		file.entry.Attributes.Crtime = req.Crtime.Unix()
+		entry.Attributes.Crtime = req.Crtime.Unix()
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Mtime() {
-		file.entry.Attributes.Mtime = req.Mtime.Unix()
+		entry.Attributes.Mtime = req.Mtime.Unix()
 		file.dirtyMetadata = true
 	}
 
@@ -188,12 +198,16 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 		return nil
 	}
 
-	return file.saveEntry(file.entry)
+	return file.saveEntry(entry)
 
 }
 
 func (file *File) Setxattr(ctx context.Context, req *fuse.SetxattrRequest) error {
 
+	if file.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("file Setxattr %s: %s", file.fullpath(), req.Name)
 
 	entry, err := file.maybeLoadEntry(ctx)
@@ -211,6 +225,10 @@ func (file *File) Setxattr(ctx context.Context, req *fuse.SetxattrRequest) error
 
 func (file *File) Removexattr(ctx context.Context, req *fuse.RemovexattrRequest) error {
 
+	if file.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	glog.V(4).Infof("file Removexattr %s: %s", file.fullpath(), req.Name)
 
 	entry, err := file.maybeLoadEntry(ctx)
@@ -255,10 +273,12 @@ func (file *File) Forget() {
 	t := util.NewFullPath(file.dir.FullPath(), file.Name)
 	glog.V(4).Infof("Forget file %s", t)
 	file.wfs.fsNodeCache.DeleteFsNode(t)
+	file.wfs.ReleaseHandle(t, 0)
+	file.setReader(nil)
 }
 
 func (file *File) maybeLoadEntry(ctx context.Context) (entry *filer_pb.Entry, err error) {
-	entry = file.entry
+	entry = file.getEntry()
 	if file.isOpen > 0 {
 		return entry, nil
 	}
@@ -299,8 +319,13 @@ func (file *File) addChunks(chunks []*filer_pb.FileChunk) {
 		}
 	}
 
+	entry := file.getEntry()
+	if entry == nil {
+		return
+	}
+
 	// pick out-of-order chunks from existing chunks
-	for _, chunk := range file.entry.Chunks {
+	for _, chunk := range entry.Chunks {
 		if lessThan(earliestChunk, chunk) {
 			chunks = append(chunks, chunk)
 		}
@@ -316,23 +341,37 @@ func (file *File) addChunks(chunks []*filer_pb.FileChunk) {
 		file.entryViewCache = filer.MergeIntoVisibles(file.entryViewCache, chunk)
 	}
 
-	file.reader = nil
+	file.setReader(nil)
 
-	glog.V(4).Infof("%s existing %d chunks adds %d more", file.fullpath(), len(file.entry.Chunks), len(chunks))
+	glog.V(4).Infof("%s existing %d chunks adds %d more", file.fullpath(), len(entry.Chunks), len(chunks))
 
-	file.entry.Chunks = append(file.entry.Chunks, newChunks...)
+	entry.Chunks = append(entry.Chunks, newChunks...)
+}
+
+func (file *File) setReader(reader io.ReaderAt) {
+	r := file.reader
+	if r != nil {
+		if closer, ok := r.(io.Closer); ok {
+			closer.Close()
+		}
+	}
+	file.reader = reader
 }
 
 func (file *File) setEntry(entry *filer_pb.Entry) {
+	file.entryLock.Lock()
+	defer file.entryLock.Unlock()
 	file.entry = entry
 	file.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(file.wfs.LookupFn(), entry.Chunks)
-	file.reader = nil
+	file.setReader(nil)
 }
 
 func (file *File) clearEntry() {
+	file.entryLock.Lock()
+	defer file.entryLock.Unlock()
 	file.entry = nil
 	file.entryViewCache = nil
-	file.reader = nil
+	file.setReader(nil)
 }
 
 func (file *File) saveEntry(entry *filer_pb.Entry) error {
@@ -359,3 +398,9 @@ func (file *File) saveEntry(entry *filer_pb.Entry) error {
 		return nil
 	})
 }
+
+func (file *File) getEntry() *filer_pb.Entry {
+	file.entryLock.RLock()
+	defer file.entryLock.RUnlock()
+	return file.entry
+}

weed/filesys/filehandle.go

@@ -23,7 +23,7 @@ type FileHandle struct {
 	dirtyPages  *ContinuousDirtyPages
 	contentType string
 	handle      uint64
-	sync.RWMutex
+	sync.Mutex
 
 	f         *File
 	RequestId fuse.RequestID // unique ID for request
@@ -40,8 +40,9 @@ func newFileHandle(file *File, uid, gid uint32) *FileHandle {
 		Uid:        uid,
 		Gid:        gid,
 	}
-	if fh.f.entry != nil {
-		fh.f.entry.Attributes.FileSize = filer.FileSize(fh.f.entry)
+	entry := fh.f.getEntry()
+	if entry != nil {
+		entry.Attributes.FileSize = filer.FileSize(entry)
 	}
 
 	return fh
@@ -58,8 +59,8 @@ var _ = fs.HandleReleaser(&FileHandle{})
 func (fh *FileHandle) Read(ctx context.Context, req *fuse.ReadRequest, resp *fuse.ReadResponse) error {
 
 	glog.V(4).Infof("%s read fh %d: [%d,%d) size %d resp.Data cap=%d", fh.f.fullpath(), fh.handle, req.Offset, req.Offset+int64(req.Size), req.Size, cap(resp.Data))
-	fh.RLock()
-	defer fh.RUnlock()
+	fh.Lock()
+	defer fh.Unlock()
 
 	if req.Size <= 0 {
 		return nil
@@ -104,26 +105,32 @@ func (fh *FileHandle) readFromDirtyPages(buff []byte, startOffset int64) (maxSto
 
 func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 
-	fileSize := int64(filer.FileSize(fh.f.entry))
-
-	if fileSize == 0 {
-		glog.V(1).Infof("empty fh %v", fh.f.fullpath())
+	entry := fh.f.getEntry()
+	if entry == nil {
 		return 0, io.EOF
 	}
 
-	if offset+int64(len(buff)) <= int64(len(fh.f.entry.Content)) {
-		totalRead := copy(buff, fh.f.entry.Content[offset:])
-		glog.V(4).Infof("file handle read cached %s [%d,%d] %d", fh.f.fullpath(), offset, offset+int64(totalRead), totalRead)
+	fileSize := int64(filer.FileSize(entry))
+	fileFullPath := fh.f.fullpath()
+
+	if fileSize == 0 {
+		glog.V(1).Infof("empty fh %v", fileFullPath)
+		return 0, io.EOF
+	}
+
+	if offset+int64(len(buff)) <= int64(len(entry.Content)) {
+		totalRead := copy(buff, entry.Content[offset:])
+		glog.V(4).Infof("file handle read cached %s [%d,%d] %d", fileFullPath, offset, offset+int64(totalRead), totalRead)
 		return int64(totalRead), nil
 	}
 
 	var chunkResolveErr error
 	if fh.f.entryViewCache == nil {
-		fh.f.entryViewCache, chunkResolveErr = filer.NonOverlappingVisibleIntervals(fh.f.wfs.LookupFn(), fh.f.entry.Chunks)
+		fh.f.entryViewCache, chunkResolveErr = filer.NonOverlappingVisibleIntervals(fh.f.wfs.LookupFn(), entry.Chunks)
 		if chunkResolveErr != nil {
 			return 0, fmt.Errorf("fail to resolve chunk manifest: %v", chunkResolveErr)
 		}
-		fh.f.reader = nil
+		fh.f.setReader(nil)
 	}
 
 	reader := fh.f.reader
@@ -131,15 +138,15 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 		chunkViews := filer.ViewFromVisibleIntervals(fh.f.entryViewCache, 0, math.MaxInt64)
 		reader = filer.NewChunkReaderAtFromClient(fh.f.wfs.LookupFn(), chunkViews, fh.f.wfs.chunkCache, fileSize)
 	}
-	fh.f.reader = reader
+	fh.f.setReader(reader)
 
 	totalRead, err := reader.ReadAt(buff, offset)
 
 	if err != nil && err != io.EOF {
-		glog.Errorf("file handle read %s: %v", fh.f.fullpath(), err)
+		glog.Errorf("file handle read %s: %v", fileFullPath, err)
 	}
 
-	glog.V(4).Infof("file handle read %s [%d,%d] %d : %v", fh.f.fullpath(), offset, offset+int64(totalRead), totalRead, err)
+	glog.V(4).Infof("file handle read %s [%d,%d] %d : %v", fileFullPath, offset, offset+int64(totalRead), totalRead, err)
 
 	return int64(totalRead), err
 }
@@ -147,6 +154,10 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 // Write to the file handle
 func (fh *FileHandle) Write(ctx context.Context, req *fuse.WriteRequest, resp *fuse.WriteResponse) error {
 
+	if fh.f.wfs.option.ReadOnly {
+		return fuse.EPERM
+	}
+
 	fh.Lock()
 	defer fh.Unlock()
 
@@ -158,8 +169,13 @@ func (fh *FileHandle) Write(ctx context.Context, req *fuse.WriteRequest, resp *f
 		copy(data, req.Data)
 	}
 
-	fh.f.entry.Content = nil
-	fh.f.entry.Attributes.FileSize = uint64(max(req.Offset+int64(len(data)), int64(fh.f.entry.Attributes.FileSize)))
+	entry := fh.f.getEntry()
+	if entry == nil {
+		return fuse.EIO
+	}
+
+	entry.Content = nil
+	entry.Attributes.FileSize = uint64(max(req.Offset+int64(len(data)), int64(entry.Attributes.FileSize)))
 	glog.V(4).Infof("%v write [%d,%d) %d", fh.f.fullpath(), req.Offset, req.Offset+int64(len(req.Data)), len(req.Data))
 
 	fh.dirtyPages.AddPage(req.Offset, data)
@@ -195,12 +211,7 @@ func (fh *FileHandle) Release(ctx context.Context, req *fuse.ReleaseRequest) err
 		fh.f.isOpen--
 
 		fh.f.wfs.ReleaseHandle(fh.f.fullpath(), fuse.HandleID(fh.handle))
-		if closer, ok := fh.f.reader.(io.Closer); ok {
-			if closer != nil {
-				closer.Close()
-			}
-		}
-		fh.f.reader = nil
+		fh.f.setReader(nil)
 	}
 
 	return nil
@@ -242,35 +253,40 @@ func (fh *FileHandle) doFlush(ctx context.Context, header fuse.Header) error {
 
 	err := fh.f.wfs.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
 
-		if fh.f.entry.Attributes != nil {
-			fh.f.entry.Attributes.Mime = fh.contentType
-			if fh.f.entry.Attributes.Uid == 0 {
-				fh.f.entry.Attributes.Uid = header.Uid
+		entry := fh.f.getEntry()
+		if entry == nil {
+			return nil
+		}
+
+		if entry.Attributes != nil {
+			entry.Attributes.Mime = fh.contentType
+			if entry.Attributes.Uid == 0 {
+				entry.Attributes.Uid = header.Uid
 			}
-			if fh.f.entry.Attributes.Gid == 0 {
-				fh.f.entry.Attributes.Gid = header.Gid
+			if entry.Attributes.Gid == 0 {
+				entry.Attributes.Gid = header.Gid
 			}
-			if fh.f.entry.Attributes.Crtime == 0 {
-				fh.f.entry.Attributes.Crtime = time.Now().Unix()
+			if entry.Attributes.Crtime == 0 {
+				entry.Attributes.Crtime = time.Now().Unix()
 			}
-			fh.f.entry.Attributes.Mtime = time.Now().Unix()
-			fh.f.entry.Attributes.FileMode = uint32(os.FileMode(fh.f.entry.Attributes.FileMode) &^ fh.f.wfs.option.Umask)
-			fh.f.entry.Attributes.Collection = fh.dirtyPages.collection
-			fh.f.entry.Attributes.Replication = fh.dirtyPages.replication
+			entry.Attributes.Mtime = time.Now().Unix()
+			entry.Attributes.FileMode = uint32(os.FileMode(entry.Attributes.FileMode) &^ fh.f.wfs.option.Umask)
+			entry.Attributes.Collection = fh.dirtyPages.collection
+			entry.Attributes.Replication = fh.dirtyPages.replication
 		}
 
 		request := &filer_pb.CreateEntryRequest{
 			Directory:  fh.f.dir.FullPath(),
-			Entry:      fh.f.entry,
+			Entry:      entry,
 			Signatures: []int32{fh.f.wfs.signature},
 		}
 
-		glog.V(4).Infof("%s set chunks: %v", fh.f.fullpath(), len(fh.f.entry.Chunks))
-		for i, chunk := range fh.f.entry.Chunks {
+		glog.V(4).Infof("%s set chunks: %v", fh.f.fullpath(), len(entry.Chunks))
+		for i, chunk := range entry.Chunks {
 			glog.V(4).Infof("%s chunks %d: %v [%d,%d)", fh.f.fullpath(), i, chunk.GetFileIdString(), chunk.Offset, chunk.Offset+int64(chunk.Size))
 		}
 
-		manifestChunks, nonManifestChunks := filer.SeparateManifestChunks(fh.f.entry.Chunks)
+		manifestChunks, nonManifestChunks := filer.SeparateManifestChunks(entry.Chunks)
 
 		chunks, _ := filer.CompactFileChunks(fh.f.wfs.LookupFn(), nonManifestChunks)
 		chunks, manifestErr := filer.MaybeManifestize(fh.f.wfs.saveDataAsChunk(fh.f.fullpath()), chunks)
@@ -278,7 +294,7 @@ func (fh *FileHandle) doFlush(ctx context.Context, header fuse.Header) error {
 			// not good, but should be ok
 			glog.V(0).Infof("MaybeManifestize: %v", manifestErr)
 		}
-		fh.f.entry.Chunks = append(chunks, manifestChunks...)
+		entry.Chunks = append(chunks, manifestChunks...)
 
 		fh.f.wfs.mapPbIdFromLocalToFiler(request.Entry)
 		defer fh.f.wfs.mapPbIdFromFilerToLocal(request.Entry)

weed/filesys/fscache.go

@@ -124,8 +124,9 @@ func (c *FsCache) Move(oldPath util.FullPath, newPath util.FullPath) *FsNode {
 	}
 	if f, ok := src.node.(*File); ok {
 		f.Name = target.name
-		if f.entry != nil {
-			f.entry.Name = f.Name
+		entry := f.getEntry()
+		if entry != nil {
+			entry.Name = f.Name
 		}
 	}
 	parent.disconnectChild(target)

weed/filesys/wfs.go

@@ -43,6 +43,7 @@ type Option struct {
 	DataCenter         string
 	EntryCacheTtl      time.Duration
 	Umask              os.FileMode
+	ReadOnly           bool
 
 	MountUid   uint32
 	MountGid   uint32

weed/operation/buffer_pool.go

@@ -0,0 +1,24 @@
+package operation
+
+import (
+	"github.com/valyala/bytebufferpool"
+	"sync/atomic"
+)
+
+var bufferCounter int64
+
+func GetBuffer() *bytebufferpool.ByteBuffer {
+	defer func() {
+		atomic.AddInt64(&bufferCounter, 1)
+		// println("+", bufferCounter)
+	}()
+	return bytebufferpool.Get()
+}
+
+func PutBuffer(buf *bytebufferpool.ByteBuffer) {
+	defer func() {
+		atomic.AddInt64(&bufferCounter, -1)
+		// println("-", bufferCounter)
+	}()
+	bytebufferpool.Put(buf)
+}

weed/operation/upload_content.go

@@ -19,7 +19,6 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
 	"github.com/chrislusf/seaweedfs/weed/security"
 	"github.com/chrislusf/seaweedfs/weed/util"
-	"github.com/valyala/bytebufferpool"
 )
 
 type UploadResult struct {
@@ -31,6 +30,7 @@ type UploadResult struct {
 	Mime       string `json:"mime,omitempty"`
 	Gzip       uint32 `json:"gzip,omitempty"`
 	ContentMd5 string `json:"contentMd5,omitempty"`
+	RetryCount int    `json:"-"`
 }
 
 func (uploadResult *UploadResult) ToPbFileChunk(fileId string, offset int64) *filer_pb.FileChunk {
@@ -96,6 +96,7 @@ func retriedUploadData(uploadUrl string, filename string, cipher bool, data []by
 	for i := 0; i < 3; i++ {
 		uploadResult, err = doUploadData(uploadUrl, filename, cipher, data, isInputCompressed, mtype, pairMap, jwt)
 		if err == nil {
+			uploadResult.RetryCount = i
 			return
 		} else {
 			glog.Warningf("uploading to %s: %v", uploadUrl, err)
@@ -130,7 +131,8 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 	// gzip if possible
 	// this could be double copying
 	clearDataLen = len(data)
-	if shouldGzipNow {
+	clearData := data
+	if shouldGzipNow && !cipher {
 		compressed, compressErr := util.GzipData(data)
 		// fmt.Printf("data is compressed from %d ==> %d\n", len(data), len(compressed))
 		if compressErr == nil {
@@ -139,7 +141,7 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 		}
 	} else if isInputCompressed {
 		// just to get the clear data length
-		clearData, err := util.DecompressData(data)
+		clearData, err = util.DecompressData(data)
 		if err == nil {
 			clearDataLen = len(clearData)
 		}
@@ -150,7 +152,7 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 
 		// encrypt
 		cipherKey := util.GenCipherKey()
-		encryptedData, encryptionErr := util.Encrypt(data, cipherKey)
+		encryptedData, encryptionErr := util.Encrypt(clearData, cipherKey)
 		if encryptionErr != nil {
 			err = fmt.Errorf("encrypt input: %v", encryptionErr)
 			return
@@ -161,34 +163,34 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 			_, err = w.Write(encryptedData)
 			return
 		}, "", false, len(encryptedData), "", nil, jwt)
-		if uploadResult != nil {
-			uploadResult.Name = filename
-			uploadResult.Mime = mtype
-			uploadResult.CipherKey = cipherKey
+		if uploadResult == nil {
+			return
 		}
+		uploadResult.Name = filename
+		uploadResult.Mime = mtype
+		uploadResult.CipherKey = cipherKey
+		uploadResult.Size = uint32(clearDataLen)
 	} else {
 		// upload data
 		uploadResult, err = upload_content(uploadUrl, func(w io.Writer) (err error) {
 			_, err = w.Write(data)
 			return
 		}, filename, contentIsGzipped, len(data), mtype, pairMap, jwt)
-	}
-
-	if uploadResult == nil {
-		return
-	}
-
-	uploadResult.Size = uint32(clearDataLen)
-	if contentIsGzipped {
-		uploadResult.Gzip = 1
+		if uploadResult == nil {
+			return
+		}
+		uploadResult.Size = uint32(clearDataLen)
+		if contentIsGzipped {
+			uploadResult.Gzip = 1
+		}
 	}
 
 	return uploadResult, err
 }
 
 func upload_content(uploadUrl string, fillBufferFunction func(w io.Writer) error, filename string, isGzipped bool, originalDataSize int, mtype string, pairMap map[string]string, jwt security.EncodedJwt) (*UploadResult, error) {
-	buf := bytebufferpool.Get()
-	defer bytebufferpool.Put(buf)
+	buf := GetBuffer()
+	defer PutBuffer(buf)
 	body_writer := multipart.NewWriter(buf)
 	h := make(textproto.MIMEHeader)
 	h.Set("Content-Disposition", fmt.Sprintf(`form-data; name="file"; filename="%s"`, fileNameEscaper.Replace(filename)))

weed/pb/filer_pb/filer.pb.go

@@ -1,6 +1,6 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
 // versions:
-// 	protoc-gen-go v1.24.0
+// 	protoc-gen-go v1.25.0
 // 	protoc        v3.12.3
 // source: filer.proto
 

weed/pb/filer_pb/filer_client.go

@@ -5,6 +5,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"math"
 	"os"
 	"strings"
 	"time"
@@ -101,12 +102,16 @@ func SeaweedList(client SeaweedFilerClient, parentDirectoryPath, prefix string,
 }
 
 func doSeaweedList(client SeaweedFilerClient, fullDirPath util.FullPath, prefix string, fn EachEntryFunciton, startFrom string, inclusive bool, limit uint32) (err error) {
-
+	// Redundancy limit to make it correctly judge whether it is the last file.
+	redLimit := limit
+	if limit != math.MaxInt32 && limit != 0 {
+		redLimit = limit + 1
+	}
 	request := &ListEntriesRequest{
 		Directory:          string(fullDirPath),
 		Prefix:             prefix,
 		StartFromFileName:  startFrom,
-		Limit:              limit,
+		Limit:              redLimit,
 		InclusiveStartFrom: inclusive,
 	}
 
@@ -119,6 +124,7 @@ func doSeaweedList(client SeaweedFilerClient, fullDirPath util.FullPath, prefix
 	}
 
 	var prevEntry *Entry
+	count := 0
 	for {
 		resp, recvErr := stream.Recv()
 		if recvErr != nil {
@@ -139,6 +145,10 @@ func doSeaweedList(client SeaweedFilerClient, fullDirPath util.FullPath, prefix
 			}
 		}
 		prevEntry = resp.Entry
+		count++
+		if count > int(limit) && limit != 0 {
+			prevEntry = nil
+		}
 	}
 
 	return nil
@@ -172,6 +182,26 @@ func Exists(filerClient FilerClient, parentDirectoryPath string, entryName strin
 	return
 }
 
+func Touch(filerClient FilerClient, parentDirectoryPath string, entryName string, entry *Entry) (err error) {
+
+	return filerClient.WithFilerClient(func(client SeaweedFilerClient) error {
+
+		request := &UpdateEntryRequest{
+			Directory: parentDirectoryPath,
+			Entry:     entry,
+		}
+
+		glog.V(4).Infof("touch entry %v/%v: %v", parentDirectoryPath, entryName, request)
+		if err := UpdateEntry(client, request); err != nil {
+			glog.V(0).Infof("touch exists entry %v: %v", request, err)
+			return fmt.Errorf("touch exists entry %s/%s: %v", parentDirectoryPath, entryName, err)
+		}
+
+		return nil
+	})
+
+}
+
 func Mkdir(filerClient FilerClient, parentDirectoryPath string, dirName string, fn func(entry *Entry)) error {
 	return filerClient.WithFilerClient(func(client SeaweedFilerClient) error {
 

weed/pb/grpc_client_server.go

@@ -3,6 +3,7 @@ package pb
 import (
 	"context"
 	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/glog"
 	"net/http"
 	"strconv"
 	"strings"
@@ -108,51 +109,55 @@ func WithCachedGrpcClient(fn func(*grpc.ClientConn) error, address string, opts
 }
 
 func ParseServerToGrpcAddress(server string) (serverGrpcAddress string, err error) {
-	colonIndex := strings.LastIndex(server, ":")
-	if colonIndex < 0 {
-		return "", fmt.Errorf("server should have hostname:port format: %v", server)
-	}
+	return ParseServerAddress(server, 10000)
+}
 
-	port, parseErr := strconv.ParseUint(server[colonIndex+1:], 10, 64)
+func ParseServerAddress(server string, deltaPort int) (newServerAddress string, err error) {
+
+	host, port, parseErr := hostAndPort(server)
 	if parseErr != nil {
 		return "", fmt.Errorf("server port parse error: %v", parseErr)
 	}
 
-	grpcPort := int(port) + 10000
+	newPort := int(port) + deltaPort
 
-	return fmt.Sprintf("%s:%d", server[:colonIndex], grpcPort), nil
+	return fmt.Sprintf("%s:%d", host, newPort), nil
+}
+
+func hostAndPort(address string) (host string, port uint64, err error) {
+	colonIndex := strings.LastIndex(address, ":")
+	if colonIndex < 0 {
+		return "", 0, fmt.Errorf("server should have hostname:port format: %v", address)
+	}
+	port, err = strconv.ParseUint(address[colonIndex+1:], 10, 64)
+	if err != nil {
+		return "", 0, fmt.Errorf("server port parse error: %v", err)
+	}
+
+	return address[:colonIndex], port, err
 }
 
 func ServerToGrpcAddress(server string) (serverGrpcAddress string) {
-	hostnameAndPort := strings.Split(server, ":")
-	if len(hostnameAndPort) != 2 {
-		return fmt.Sprintf("unexpected server address: %s", server)
-	}
 
-	port, parseErr := strconv.ParseUint(hostnameAndPort[1], 10, 64)
+	host, port, parseErr := hostAndPort(server)
 	if parseErr != nil {
-		return fmt.Sprintf("failed to parse port for %s:%s", hostnameAndPort[0], hostnameAndPort[1])
+		glog.Fatalf("server address %s parse error: %v", server, parseErr)
 	}
 
 	grpcPort := int(port) + 10000
 
-	return fmt.Sprintf("%s:%d", hostnameAndPort[0], grpcPort)
+	return fmt.Sprintf("%s:%d", host, grpcPort)
 }
 
 func GrpcAddressToServerAddress(grpcAddress string) (serverAddress string) {
-	hostnameAndPort := strings.Split(grpcAddress, ":")
-	if len(hostnameAndPort) != 2 {
-		return fmt.Sprintf("unexpected grpcAddress: %s", grpcAddress)
-	}
-
-	grpcPort, parseErr := strconv.ParseUint(hostnameAndPort[1], 10, 64)
+	host, grpcPort, parseErr := hostAndPort(grpcAddress)
 	if parseErr != nil {
-		return fmt.Sprintf("failed to parse port for %s:%s", hostnameAndPort[0], hostnameAndPort[1])
+		glog.Fatalf("server grpc address %s parse error: %v", grpcAddress, parseErr)
 	}
 
 	port := int(grpcPort) - 10000
 
-	return fmt.Sprintf("%s:%d", hostnameAndPort[0], port)
+	return fmt.Sprintf("%s:%d", host, port)
 }
 
 func WithMasterClient(master string, grpcDialOption grpc.DialOption, fn func(client master_pb.SeaweedClient) error) error {
@@ -197,19 +202,3 @@ func WithGrpcFilerClient(filerGrpcAddress string, grpcDialOption grpc.DialOption
 	}, filerGrpcAddress, grpcDialOption)
 
 }
-
-func ParseFilerGrpcAddress(filer string) (filerGrpcAddress string, err error) {
-	hostnameAndPort := strings.Split(filer, ":")
-	if len(hostnameAndPort) != 2 {
-		return "", fmt.Errorf("filer should have hostname:port format: %v", hostnameAndPort)
-	}
-
-	filerPort, parseErr := strconv.ParseUint(hostnameAndPort[1], 10, 64)
-	if parseErr != nil {
-		return "", fmt.Errorf("filer port parse error: %v", parseErr)
-	}
-
-	filerGrpcPort := int(filerPort) + 10000
-
-	return fmt.Sprintf("%s:%d", hostnameAndPort[0], filerGrpcPort), nil
-}

weed/pb/iam_pb/iam.pb.go

@@ -1,6 +1,6 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
 // versions:
-// 	protoc-gen-go v1.24.0
+// 	protoc-gen-go v1.25.0
 // 	protoc        v3.12.3
 // source: iam.proto
 

weed/pb/master_pb/master.pb.go

@@ -1,6 +1,6 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
 // versions:
-// 	protoc-gen-go v1.24.0
+// 	protoc-gen-go v1.25.0
 // 	protoc        v3.12.3
 // source: master.proto
 

weed/pb/messaging_pb/messaging.pb.go

@@ -1,6 +1,6 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
 // versions:
-// 	protoc-gen-go v1.24.0
+// 	protoc-gen-go v1.25.0
 // 	protoc        v3.12.3
 // source: messaging.proto
 

weed/pb/volume_info.go

@@ -15,40 +15,49 @@ import (
 )
 
 // MaybeLoadVolumeInfo load the file data as *volume_server_pb.VolumeInfo, the returned volumeInfo will not be nil
-func MaybeLoadVolumeInfo(fileName string) (*volume_server_pb.VolumeInfo, bool, error) {
+func MaybeLoadVolumeInfo(fileName string) (volumeInfo *volume_server_pb.VolumeInfo, hasRemoteFile bool, hasVolumeInfoFile bool, err error) {
 
-	volumeInfo := &volume_server_pb.VolumeInfo{}
+	volumeInfo = &volume_server_pb.VolumeInfo{}
 
 	glog.V(1).Infof("maybeLoadVolumeInfo checks %s", fileName)
 	if exists, canRead, _, _, _ := util.CheckFile(fileName); !exists || !canRead {
 		if !exists {
-			return volumeInfo, false, nil
+			return
 		}
+		hasVolumeInfoFile = true
 		if !canRead {
 			glog.Warningf("can not read %s", fileName)
-			return volumeInfo, false, fmt.Errorf("can not read %s", fileName)
+			err = fmt.Errorf("can not read %s", fileName)
+			return
 		}
-		return volumeInfo, false, nil
+		return
 	}
 
+	hasVolumeInfoFile = true
+
 	glog.V(1).Infof("maybeLoadVolumeInfo reads %s", fileName)
 	tierData, readErr := ioutil.ReadFile(fileName)
 	if readErr != nil {
 		glog.Warningf("fail to read %s : %v", fileName, readErr)
-		return volumeInfo, false, fmt.Errorf("fail to read %s : %v", fileName, readErr)
+		err = fmt.Errorf("fail to read %s : %v", fileName, readErr)
+		return
+
 	}
 
 	glog.V(1).Infof("maybeLoadVolumeInfo Unmarshal volume info %v", fileName)
-	if err := jsonpb.Unmarshal(bytes.NewReader(tierData), volumeInfo); err != nil {
+	if err = jsonpb.Unmarshal(bytes.NewReader(tierData), volumeInfo); err != nil {
 		glog.Warningf("unmarshal error: %v", err)
-		return volumeInfo, false, fmt.Errorf("unmarshal error: %v", err)
+		err = fmt.Errorf("unmarshal error: %v", err)
+		return
 	}
 
 	if len(volumeInfo.GetFiles()) == 0 {
-		return volumeInfo, false, nil
+		return
 	}
 
-	return volumeInfo, true, nil
+	hasRemoteFile = true
+
+	return
 }
 
 func SaveVolumeInfo(fileName string, volumeInfo *volume_server_pb.VolumeInfo) error {