fix to avoid loop

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

avoid shared readErr variable

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 ##
 

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-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_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|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

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
@@ -38,7 +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 // 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
@@ -88,7 +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=
@@ -959,6 +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/seaweedfs/Chart.yaml

@@ -1,5 +1,5 @@
 apiVersion: v1
 description: SeaweedFS
 name: seaweedfs
-appVersion: "2.32"
-version: 2.32
+appVersion: "2.38"
+version: 2.38

k8s/seaweedfs/values.yaml

@@ -4,7 +4,7 @@ global:
   registry: ""
   repository: ""
   imageName: chrislusf/seaweedfs
-  # imageTag: "2.32" - started using {.Chart.appVersion}
+  # imageTag: "2.38" - started using {.Chart.appVersion}
   imagePullPolicy: IfNotPresent
   imagePullSecrets: imagepullsecret
   restartPolicy: Always

other/java/client/pom.xml

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.6.2</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>1.6.2</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>1.6.2</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/examples/pom.xml

@@ -11,13 +11,13 @@
         <dependency>
             <groupId>com.github.chrislusf</groupId>
             <artifactId>seaweedfs-client</artifactId>
-            <version>1.6.2</version>
+            <version>1.6.4</version>
             <scope>compile</scope>
         </dependency>
         <dependency>
             <groupId>com.github.chrislusf</groupId>
             <artifactId>seaweedfs-hadoop2-client</artifactId>
-            <version>1.6.2</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>1.6.2</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>1.6.2</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>1.6.2</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>1.6.2</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/command/benchmark.go

@@ -305,7 +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
 			}

weed/command/filer.go

@@ -47,6 +47,7 @@ type FilerOptions struct {
 	metricsHttpPort         *int
 	saveToFilerLimit        *int
 	defaultLevelDbDirectory *string
+	concurrentUploadLimitMB *int
 }
 
 func init() {
@@ -59,7 +60,7 @@ func init() {
 	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")
@@ -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_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

@@ -56,7 +56,7 @@ func init() {
 	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|<tag>] hard drive or solid state drive or any tag")
-	copy.maxMB = cmdCopy.Flag.Int("maxMB", 32, "split files larger than the limit")
+	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/master.go

@@ -6,7 +6,6 @@ import (
 	"google.golang.org/grpc/reflection"
 	"net/http"
 	"os"
-	"runtime"
 	"sort"
 	"strconv"
 	"strings"
@@ -48,7 +47,7 @@ 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.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")
@@ -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()

weed/command/mount.go

@@ -25,6 +25,7 @@ type MountOptions struct {
 	volumeServerAccess *string
 	uidMap             *string
 	gidMap             *string
+	readOnly           *bool
 }
 
 var (
@@ -45,7 +46,7 @@ func init() {
 	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

@@ -203,6 +203,7 @@ func RunMount(option *MountOptions, umask os.FileMode) bool {
 		VolumeServerAccess: *mountOptions.volumeServerAccess,
 		Cipher:             cipher,
 		UidGidMapper:       uidGidMapper,
+		ReadOnly:           *option.readOnly,
 	})
 
 	// mount

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 (
@@ -506,7 +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
 }
 
@@ -49,7 +49,7 @@ var cmdServer = &Command{
 }
 
 var (
-	serverIp                  = cmdServer.Flag.String("ip", util.DetectedHostAddress(), "ip or server name")
+	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")
@@ -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")
@@ -93,11 +94,12 @@ 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")
@@ -107,6 +109,7 @@ func init() {
 	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")
@@ -138,14 +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
@@ -192,7 +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

@@ -43,7 +43,7 @@ func init() {
 	upload.dataCenter = cmdUpload.Flag.String("dataCenter", "", "optional data center name")
 	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"
@@ -36,31 +35,32 @@ 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
 }
@@ -69,7 +69,7 @@ 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", "", "ip address to bind to")
 	v.masters = cmdVolume.Flag.String("mserver", "localhost:9333", "comma-separated master servers")
@@ -86,6 +86,7 @@ func init() {
 	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")
@@ -111,8 +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 {
@@ -240,6 +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)

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 {
@@ -36,6 +36,8 @@ func (store *AbstractSqlStore) OnBucketCreation(bucket string) {
 	store.dbsLock.Lock()
 	defer store.dbsLock.Unlock()
 
+	store.CreateTable(context.Background(), bucket)
+
 	if store.dbs == nil {
 		return
 	}
@@ -45,6 +47,8 @@ func (store *AbstractSqlStore) OnBucketDeletion(bucket string) {
 	store.dbsLock.Lock()
 	defer store.dbsLock.Unlock()
 
+	store.deleteTable(context.Background(), bucket)
+
 	if store.dbs == nil {
 		return
 	}

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/filechunks.go

@@ -23,10 +23,6 @@ func TotalSize(chunks []*filer_pb.FileChunk) (size uint64) {
 	return
 }
 
-func FileSize2(entry *Entry) (size uint64) {
-	return maxUint64(TotalSize(entry.Chunks), entry.Attr.FileSize)
-}
-
 func FileSize(entry *filer_pb.Entry) (size uint64) {
 	return maxUint64(TotalSize(entry.Chunks), entry.Attributes.FileSize)
 }

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_on_meta_event.go

@@ -52,7 +52,7 @@ func (f *Filer) maybeReloadFilerConfiguration(event *filer_pb.SubscribeMetadataR
 
 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/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/stream.go

@@ -3,6 +3,7 @@ package filer
 import (
 	"bytes"
 	"fmt"
+	"golang.org/x/sync/errgroup"
 	"io"
 	"math"
 	"strings"
@@ -13,7 +14,7 @@ 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 {
 
 	glog.V(9).Infof("start to stream content for chunks: %+v\n", chunks)
 	chunkViews := ViewFromChunks(masterClient.GetLookupFileIdFunction(), chunks, offset, size)
@@ -33,10 +34,20 @@ func StreamContent(masterClient wdclient.HasLookupFileIdFunction, w io.Writer, c
 		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)
@@ -54,6 +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) {
@@ -185,7 +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, 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

@@ -1,6 +1,7 @@
 package filesys
 
 import (
+	"bytes"
 	"context"
 	"math"
 	"os"
@@ -21,7 +22,7 @@ import (
 type Dir struct {
 	name   string
 	wfs    *WFS
-	entry  *filer.Entry
+	entry  *filer_pb.Entry
 	parent *Dir
 }
 
@@ -58,11 +59,11 @@ func (dir *Dir) Attr(ctx context.Context, attr *fuse.Attr) error {
 	}
 
 	// attr.Inode = util.FullPath(dir.FullPath()).AsInode()
-	attr.Mode = dir.entry.Attr.Mode | os.ModeDir
-	attr.Mtime = dir.entry.Attr.Mtime
-	attr.Crtime = dir.entry.Attr.Crtime
-	attr.Gid = dir.entry.Attr.Gid
-	attr.Uid = dir.entry.Attr.Uid
+	attr.Mode = os.FileMode(dir.entry.Attributes.FileMode) | os.ModeDir
+	attr.Mtime = time.Unix(dir.entry.Attributes.Mtime, 0)
+	attr.Crtime = time.Unix(dir.entry.Attributes.Crtime, 0)
+	attr.Gid = dir.entry.Attributes.Gid
+	attr.Uid = dir.entry.Attributes.Uid
 
 	glog.V(4).Infof("dir Attr %s, attr: %+v", dir.FullPath(), attr)
 
@@ -102,13 +103,12 @@ func (dir *Dir) Fsync(ctx context.Context, req *fuse.FsyncRequest) error {
 }
 
 func (dir *Dir) newFile(name string, entry *filer_pb.Entry) fs.Node {
-	dirPath := dir.FullPath()
-	f := dir.wfs.fsNodeCache.EnsureFsNode(util.NewFullPath(dirPath, name), func() fs.Node {
+	f := dir.wfs.fsNodeCache.EnsureFsNode(util.NewFullPath(dir.FullPath(), name), func() fs.Node {
 		return &File{
 			Name:           name,
 			dir:            dir,
 			wfs:            dir.wfs,
-			entry:          filer.FromPbEntry(dirPath, entry),
+			entry:          entry,
 			entryViewCache: nil,
 		}
 	})
@@ -119,7 +119,7 @@ func (dir *Dir) newFile(name string, entry *filer_pb.Entry) fs.Node {
 func (dir *Dir) newDirectory(fullpath util.FullPath, entry *filer_pb.Entry) fs.Node {
 
 	d := dir.wfs.fsNodeCache.EnsureFsNode(fullpath, func() fs.Node {
-		return &Dir{name: entry.Name, wfs: dir.wfs, entry: filer.FromPbEntry(dir.FullPath(), entry), parent: dir}
+		return &Dir{name: entry.Name, wfs: dir.wfs, entry: entry, parent: dir}
 	})
 	d.(*Dir).parent = dir // in case dir node was created later
 	return d
@@ -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{
@@ -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)
 	}
@@ -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 {
@@ -436,19 +457,19 @@ func (dir *Dir) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *fus
 	}
 
 	if req.Valid.Mode() {
-		dir.entry.Attr.Mode = req.Mode
+		dir.entry.Attributes.FileMode = uint32(req.Mode)
 	}
 
 	if req.Valid.Uid() {
-		dir.entry.Attr.Uid = req.Uid
+		dir.entry.Attributes.Uid = req.Uid
 	}
 
 	if req.Valid.Gid() {
-		dir.entry.Attr.Gid = req.Gid
+		dir.entry.Attributes.Gid = req.Gid
 	}
 
 	if req.Valid.Mtime() {
-		dir.entry.Attr.Mtime = req.Mtime
+		dir.entry.Attributes.Mtime = req.Mtime.Unix()
 	}
 
 	return dir.saveEntry()
@@ -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 {
@@ -527,14 +556,12 @@ func (dir *Dir) saveEntry() error {
 
 	return dir.wfs.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
 
-		pbEntry := dir.entry.ToProtoEntry()
-
-		dir.wfs.mapPbIdFromLocalToFiler(pbEntry)
-		defer dir.wfs.mapPbIdFromFilerToLocal(pbEntry)
+		dir.wfs.mapPbIdFromLocalToFiler(dir.entry)
+		defer dir.wfs.mapPbIdFromFilerToLocal(dir.entry)
 
 		request := &filer_pb.UpdateEntryRequest{
 			Directory:  parentDir,
-			Entry:      pbEntry,
+			Entry:      dir.entry,
 			Signatures: []int32{dir.wfs.signature},
 		}
 
@@ -552,5 +579,25 @@ func (dir *Dir) saveEntry() error {
 }
 
 func (dir *Dir) FullPath() string {
-	return string(dir.entry.FullPath)
+	var parts []string
+	for p := dir; p != nil; p = p.parent {
+		if strings.HasPrefix(p.name, "/") {
+			if len(p.name) > 1 {
+				parts = append(parts, p.name[1:])
+			}
+		} else {
+			parts = append(parts, p.name)
+		}
+	}
+
+	if len(parts) == 0 {
+		return "/"
+	}
+
+	var buf bytes.Buffer
+	for i := len(parts) - 1; i >= 0; i-- {
+		buf.WriteString("/")
+		buf.WriteString(parts[i])
+	}
+	return buf.String()
 }

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)
@@ -48,7 +52,7 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 	oldEntry.HardLinkCounter++
 	updateOldEntryRequest := &filer_pb.UpdateEntryRequest{
 		Directory:  oldFile.dir.FullPath(),
-		Entry:      oldEntry.ToProtoEntry(),
+		Entry:      oldEntry,
 		Signatures: []int32{dir.wfs.signature},
 	}
 
@@ -58,7 +62,7 @@ func (dir *Dir) Link(ctx context.Context, req *fuse.LinkRequest, old fs.Node) (f
 		Entry: &filer_pb.Entry{
 			Name:            req.NewName,
 			IsDirectory:     false,
-			Attributes:      filer.EntryAttributeToPb(oldEntry),
+			Attributes:      oldEntry.Attributes,
 			Chunks:          oldEntry.Chunks,
 			Extended:        oldEntry.Extended,
 			HardLinkId:      oldEntry.HardLinkId,
@@ -105,6 +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{
@@ -152,12 +160,12 @@ func (file *File) Readlink(ctx context.Context, req *fuse.ReadlinkRequest) (stri
 		return "", err
 	}
 
-	if entry.Attr.Mode&os.ModeSymlink == 0 {
+	if os.FileMode(entry.Attributes.FileMode)&os.ModeSymlink == 0 {
 		return "", fuse.Errno(syscall.EINVAL)
 	}
 
-	glog.V(4).Infof("Readlink: %v/%v => %v", file.dir.FullPath(), file.Name, entry.Attr.SymlinkTarget)
+	glog.V(4).Infof("Readlink: %v/%v => %v", file.dir.FullPath(), file.Name, entry.Attributes.SymlinkTarget)
 
-	return entry.Attr.SymlinkTarget, nil
+	return entry.Attributes.SymlinkTarget, nil
 
 }

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/dirty_page.go

@@ -74,7 +74,7 @@ func (pages *ContinuousDirtyPages) saveExistingLargestPageToStorage() (hasSavedD
 		return false
 	}
 
-	fileSize := int64(entry.Attr.FileSize)
+	fileSize := int64(entry.Attributes.FileSize)
 
 	chunkSize := min(maxList.Size(), fileSize-maxList.Offset())
 	if chunkSize == 0 {

weed/filesys/file.go

@@ -33,7 +33,7 @@ type File struct {
 	Name           string
 	dir            *Dir
 	wfs            *WFS
-	entry          *filer.Entry
+	entry          *filer_pb.Entry
 	entryLock      sync.RWMutex
 	entryViewCache []filer.VisibleInterval
 	isOpen         int
@@ -56,18 +56,22 @@ func (file *File) Attr(ctx context.Context, attr *fuse.Attr) (err error) {
 		}
 	}
 
+	if entry == nil {
+		return fuse.ENOENT
+	}
+
 	// attr.Inode = file.fullpath().AsInode()
 	attr.Valid = time.Second
-	attr.Mode = os.FileMode(entry.Attr.Mode)
-	attr.Size = filer.FileSize2(entry)
+	attr.Mode = os.FileMode(entry.Attributes.FileMode)
+	attr.Size = filer.FileSize(entry)
 	if file.isOpen > 0 {
-		attr.Size = entry.Attr.FileSize
+		attr.Size = entry.Attributes.FileSize
 		glog.V(4).Infof("file Attr %s, open:%v, size: %d", file.fullpath(), file.isOpen, attr.Size)
 	}
-	attr.Crtime = entry.Attr.Crtime
-	attr.Mtime = entry.Attr.Mtime
-	attr.Gid = entry.Attr.Gid
-	attr.Uid = entry.Attr.Uid
+	attr.Crtime = time.Unix(entry.Attributes.Crtime, 0)
+	attr.Mtime = time.Unix(entry.Attributes.Mtime, 0)
+	attr.Gid = entry.Attributes.Gid
+	attr.Uid = entry.Attributes.Uid
 	attr.Blocks = attr.Size/blockSize + 1
 	attr.BlockSize = uint32(file.wfs.option.ChunkSizeLimit)
 	if entry.HardLinkCounter > 0 {
@@ -106,6 +110,10 @@ 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)
 
 	entry, err := file.maybeLoadEntry(ctx)
@@ -126,7 +134,7 @@ 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(entry.Chunks))
-		if req.Size < filer.FileSize2(entry) {
+		if req.Size < filer.FileSize(entry) {
 			// fmt.Printf("truncate %v \n", fullPath)
 			var chunks []*filer_pb.FileChunk
 			var truncatedChunks []*filer_pb.FileChunk
@@ -147,34 +155,34 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 			}
 			entry.Chunks = chunks
 			file.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(file.wfs.LookupFn(), chunks)
-			file.reader = nil
+			file.setReader(nil)
 		}
-		entry.Attr.FileSize = req.Size
+		entry.Attributes.FileSize = req.Size
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Mode() {
-		entry.Attr.Mode = req.Mode
+		entry.Attributes.FileMode = uint32(req.Mode)
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Uid() {
-		entry.Attr.Uid = req.Uid
+		entry.Attributes.Uid = req.Uid
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Gid() {
-		entry.Attr.Gid = req.Gid
+		entry.Attributes.Gid = req.Gid
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Crtime() {
-		entry.Attr.Crtime = req.Crtime
+		entry.Attributes.Crtime = req.Crtime.Unix()
 		file.dirtyMetadata = true
 	}
 
 	if req.Valid.Mtime() {
-		entry.Attr.Mtime = req.Mtime
+		entry.Attributes.Mtime = req.Mtime.Unix()
 		file.dirtyMetadata = true
 	}
 
@@ -196,6 +204,10 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 
 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)
@@ -213,6 +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)
@@ -257,9 +273,11 @@ 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.Entry, err error) {
+func (file *File) maybeLoadEntry(ctx context.Context) (entry *filer_pb.Entry, err error) {
 	entry = file.getEntry()
 	if file.isOpen > 0 {
 		return entry, nil
@@ -323,19 +341,29 @@ 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(entry.Chunks), len(chunks))
 
 	entry.Chunks = append(entry.Chunks, newChunks...)
 }
 
-func (file *File) setEntry(entry *filer.Entry) {
+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() {
@@ -343,20 +371,18 @@ func (file *File) clearEntry() {
 	defer file.entryLock.Unlock()
 	file.entry = nil
 	file.entryViewCache = nil
-	file.reader = nil
+	file.setReader(nil)
 }
 
-func (file *File) saveEntry(entry *filer.Entry) error {
+func (file *File) saveEntry(entry *filer_pb.Entry) error {
 	return file.wfs.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
 
-		pbEntry := entry.ToProtoEntry()
-
-		file.wfs.mapPbIdFromLocalToFiler(pbEntry)
-		defer file.wfs.mapPbIdFromFilerToLocal(pbEntry)
+		file.wfs.mapPbIdFromLocalToFiler(entry)
+		defer file.wfs.mapPbIdFromFilerToLocal(entry)
 
 		request := &filer_pb.UpdateEntryRequest{
 			Directory:  file.dir.FullPath(),
-			Entry:      pbEntry,
+			Entry:      entry,
 			Signatures: []int32{file.wfs.signature},
 		}
 
@@ -373,7 +399,7 @@ func (file *File) saveEntry(entry *filer.Entry) error {
 	})
 }
 
-func (file *File) getEntry() *filer.Entry {
+func (file *File) getEntry() *filer_pb.Entry {
 	file.entryLock.RLock()
 	defer file.entryLock.RUnlock()
 	return file.entry

weed/filesys/filehandle.go

@@ -6,6 +6,7 @@ import (
 	"io"
 	"math"
 	"net/http"
+	"os"
 	"sync"
 	"time"
 
@@ -22,7 +23,7 @@ type FileHandle struct {
 	dirtyPages  *ContinuousDirtyPages
 	contentType string
 	handle      uint64
-	sync.RWMutex
+	sync.Mutex
 
 	f         *File
 	RequestId fuse.RequestID // unique ID for request
@@ -41,7 +42,7 @@ func newFileHandle(file *File, uid, gid uint32) *FileHandle {
 	}
 	entry := fh.f.getEntry()
 	if entry != nil {
-		entry.Attr.FileSize = filer.FileSize2(entry)
+		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
@@ -109,7 +110,7 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 		return 0, io.EOF
 	}
 
-	fileSize := int64(filer.FileSize2(entry))
+	fileSize := int64(filer.FileSize(entry))
 	fileFullPath := fh.f.fullpath()
 
 	if fileSize == 0 {
@@ -129,7 +130,7 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 		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
@@ -137,7 +138,7 @@ 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)
 
@@ -153,6 +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()
 
@@ -170,7 +175,7 @@ func (fh *FileHandle) Write(ctx context.Context, req *fuse.WriteRequest, resp *f
 	}
 
 	entry.Content = nil
-	entry.Attr.FileSize = uint64(max(req.Offset+int64(len(data)), int64(entry.Attr.FileSize)))
+	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)
@@ -206,12 +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
@@ -258,24 +258,26 @@ func (fh *FileHandle) doFlush(ctx context.Context, header fuse.Header) error {
 			return nil
 		}
 
-		entry.Attr.Mime = fh.contentType
-		if entry.Attr.Uid == 0 {
-			entry.Attr.Uid = header.Uid
+		if entry.Attributes != nil {
+			entry.Attributes.Mime = fh.contentType
+			if entry.Attributes.Uid == 0 {
+				entry.Attributes.Uid = header.Uid
+			}
+			if entry.Attributes.Gid == 0 {
+				entry.Attributes.Gid = header.Gid
+			}
+			if entry.Attributes.Crtime == 0 {
+				entry.Attributes.Crtime = time.Now().Unix()
+			}
+			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
 		}
-		if entry.Attr.Gid == 0 {
-			entry.Attr.Gid = header.Gid
-		}
-		if entry.Attr.Crtime.IsZero() {
-			entry.Attr.Crtime = time.Now()
-		}
-		entry.Attr.Mtime = time.Now()
-		entry.Attr.Mode = entry.Attr.Mode &^ fh.f.wfs.option.Umask
-		entry.Attr.Collection = fh.dirtyPages.collection
-		entry.Attr.Replication = fh.dirtyPages.replication
 
 		request := &filer_pb.CreateEntryRequest{
 			Directory:  fh.f.dir.FullPath(),
-			Entry:      entry.ToProtoEntry(),
+			Entry:      entry,
 			Signatures: []int32{fh.f.wfs.signature},
 		}
 

weed/filesys/fscache.go

@@ -126,7 +126,7 @@ func (c *FsCache) Move(oldPath util.FullPath, newPath util.FullPath) *FsNode {
 		f.Name = target.name
 		entry := f.getEntry()
 		if entry != nil {
-			entry.FullPath = newPath
+			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
@@ -131,7 +132,7 @@ func NewSeaweedFileSystem(option *Option) *WFS {
 	})
 
 	entry, _ := filer_pb.GetEntry(wfs, util.FullPath(wfs.option.FilerMountRootPath))
-	wfs.root = &Dir{name: wfs.option.FilerMountRootPath, wfs: wfs, entry: filer.FromPbEntry(wfs.option.FilerMountRootPath, entry)}
+	wfs.root = &Dir{name: wfs.option.FilerMountRootPath, wfs: wfs, entry: entry}
 	wfs.fsNodeCache = newFsCache(wfs.root)
 
 	if wfs.option.ConcurrentWriters > 0 {

weed/filesys/xattr.go

@@ -2,7 +2,6 @@ package filesys
 
 import (
 	"context"
-	"github.com/chrislusf/seaweedfs/weed/filer"
 
 	"github.com/seaweedfs/fuse"
 
@@ -11,7 +10,7 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/util"
 )
 
-func getxattr(entry *filer.Entry, req *fuse.GetxattrRequest, resp *fuse.GetxattrResponse) error {
+func getxattr(entry *filer_pb.Entry, req *fuse.GetxattrRequest, resp *fuse.GetxattrResponse) error {
 
 	if entry == nil {
 		return fuse.ErrNoXattr
@@ -39,7 +38,7 @@ func getxattr(entry *filer.Entry, req *fuse.GetxattrRequest, resp *fuse.Getxattr
 
 }
 
-func setxattr(entry *filer.Entry, req *fuse.SetxattrRequest) error {
+func setxattr(entry *filer_pb.Entry, req *fuse.SetxattrRequest) error {
 
 	if entry == nil {
 		return fuse.EIO
@@ -62,7 +61,7 @@ func setxattr(entry *filer.Entry, req *fuse.SetxattrRequest) error {
 
 }
 
-func removexattr(entry *filer.Entry, req *fuse.RemovexattrRequest) error {
+func removexattr(entry *filer_pb.Entry, req *fuse.RemovexattrRequest) error {
 
 	if entry == nil {
 		return fuse.ErrNoXattr
@@ -84,7 +83,7 @@ func removexattr(entry *filer.Entry, req *fuse.RemovexattrRequest) error {
 
 }
 
-func listxattr(entry *filer.Entry, req *fuse.ListxattrRequest, resp *fuse.ListxattrResponse) error {
+func listxattr(entry *filer_pb.Entry, req *fuse.ListxattrRequest, resp *fuse.ListxattrResponse) error {
 
 	if entry == nil {
 		return fuse.EIO
@@ -109,7 +108,7 @@ func listxattr(entry *filer.Entry, req *fuse.ListxattrRequest, resp *fuse.Listxa
 
 }
 
-func (wfs *WFS) maybeLoadEntry(dir, name string) (entry *filer.Entry, err error) {
+func (wfs *WFS) maybeLoadEntry(dir, name string) (entry *filer_pb.Entry, err error) {
 
 	fullpath := util.NewFullPath(dir, name)
 	// glog.V(3).Infof("read entry cache miss %s", fullpath)
@@ -120,5 +119,5 @@ func (wfs *WFS) maybeLoadEntry(dir, name string) (entry *filer.Entry, err error)
 	if cacheErr == filer_pb.ErrNotFound {
 		return nil, fuse.ENOENT
 	}
-	return cachedEntry, cacheErr
+	return cachedEntry.ToProtoEntry(), cacheErr
 }

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 {
@@ -190,8 +189,8 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 }
 
 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

@@ -182,6 +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/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_server.proto

@@ -52,6 +52,11 @@ service VolumeServer {
     rpc CopyFile (CopyFileRequest) returns (stream CopyFileResponse) {
     }
 
+    rpc ReadNeedleBlob (ReadNeedleBlobRequest) returns (ReadNeedleBlobResponse) {
+    }
+    rpc WriteNeedleBlob (WriteNeedleBlobRequest) returns (WriteNeedleBlobResponse) {
+    }
+
     rpc VolumeTailSender (VolumeTailSenderRequest) returns (stream VolumeTailSenderResponse) {
     }
     rpc VolumeTailReceiver (VolumeTailReceiverRequest) returns (VolumeTailReceiverResponse) {
@@ -253,6 +258,25 @@ message CopyFileResponse {
     bytes file_content = 1;
 }
 
+message ReadNeedleBlobRequest {
+    uint32 volume_id = 1;
+    uint64 needle_id = 2;
+    int64 offset = 3; // actual offset
+    int32 size = 4;
+}
+message ReadNeedleBlobResponse {
+    bytes needle_blob = 1;
+}
+
+message WriteNeedleBlobRequest {
+    uint32 volume_id = 1;
+    uint64 needle_id = 2;
+    int32 size = 3;
+    bytes needle_blob = 4;
+}
+message WriteNeedleBlobResponse {
+}
+
 message VolumeTailSenderRequest {
     uint32 volume_id = 1;
     uint64 since_ns = 2;

weed/replication/repl_util/replication_util.go

@@ -20,7 +20,7 @@ func CopyFromChunkViews(chunkViews []*filer.ChunkView, filerSource *source.Filer
 		var shouldRetry bool
 
 		for _, fileUrl := range fileUrls {
-			shouldRetry, err = util.FastReadUrlAsStream(fileUrl, nil, false, chunk.IsFullChunk(), chunk.Offset, int(chunk.Size), func(data []byte) {
+			shouldRetry, err = util.ReadUrlAsStream(fileUrl, nil, false, chunk.IsFullChunk(), chunk.Offset, int(chunk.Size), func(data []byte) {
 				writeErr = writeFunc(data)
 			})
 			if err != nil {

weed/s3api/filer_util.go

@@ -79,6 +79,12 @@ func (s3a *S3ApiServer) exists(parentDirectoryPath string, entryName string, isD
 
 }
 
+func (s3a *S3ApiServer) touch(parentDirectoryPath string, entryName string, entry *filer_pb.Entry) (err error) {
+
+	return filer_pb.Touch(s3a, parentDirectoryPath, entryName, entry)
+
+}
+
 func (s3a *S3ApiServer) getEntry(parentDirectoryPath, entryName string) (entry *filer_pb.Entry, err error) {
 	fullPath := util.NewFullPath(parentDirectoryPath, entryName)
 	return filer_pb.GetEntry(s3a, fullPath)

weed/s3api/s3api_object_copy_handlers.go

@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
+	weed_server "github.com/chrislusf/seaweedfs/weed/server"
 	"net/http"
 	"net/url"
 	"strconv"
@@ -25,6 +26,26 @@ func (s3a *S3ApiServer) CopyObjectHandler(w http.ResponseWriter, r *http.Request
 	}
 
 	srcBucket, srcObject := pathToBucketAndObject(cpSrcPath)
+
+	if (srcBucket == dstBucket && srcObject == dstObject || cpSrcPath == "") && isReplace(r) {
+		fullPath := util.FullPath(fmt.Sprintf("%s/%s%s", s3a.option.BucketsPath, dstBucket, dstObject))
+		dir, name := fullPath.DirAndName()
+		entry, err := s3a.getEntry(dir, name)
+		if err != nil {
+			writeErrorResponse(w, s3err.ErrInvalidCopySource, r.URL)
+		}
+		entry.Extended = weed_server.SaveAmzMetaData(r, entry.Extended, isReplace(r))
+		err = s3a.touch(dir, name, entry)
+		if err != nil {
+			writeErrorResponse(w, s3err.ErrInvalidCopySource, r.URL)
+		}
+		writeSuccessResponseXML(w, encodeResponse(CopyObjectResult{
+			ETag:         fmt.Sprintf("%x", entry.Attributes.Md5),
+			LastModified: time.Now().UTC(),
+		}))
+		return
+	}
+
 	// If source object is empty or bucket is empty, reply back invalid copy source.
 	if srcObject == "" || srcBucket == "" {
 		writeErrorResponse(w, s3err.ErrInvalidCopySource, r.URL)
@@ -32,7 +53,7 @@ func (s3a *S3ApiServer) CopyObjectHandler(w http.ResponseWriter, r *http.Request
 	}
 
 	if srcBucket == dstBucket && srcObject == dstObject {
-		writeErrorResponse(w, s3err.ErrInvalidCopySource, r.URL)
+		writeErrorResponse(w, s3err.ErrInvalidCopyDest, r.URL)
 		return
 	}
 
@@ -147,3 +168,7 @@ func (s3a *S3ApiServer) CopyObjectPartHandler(w http.ResponseWriter, r *http.Req
 	writeSuccessResponseXML(w, encodeResponse(response))
 
 }
+
+func isReplace(r *http.Request) bool {
+	return r.Header.Get("X-Amz-Metadata-Directive") == "REPLACE"
+}

weed/s3api/s3api_object_handlers.go

@@ -311,7 +311,7 @@ func (s3a *S3ApiServer) proxyToFiler(w http.ResponseWriter, r *http.Request, des
 	}
 	defer util.CloseResponse(resp)
 
-	if (resp.ContentLength == -1 || resp.StatusCode == 404) && !strings.HasSuffix(destUrl, "/") {
+	if resp.ContentLength == -1 || resp.StatusCode == 404 {
 		if r.Method != "DELETE" {
 			writeErrorResponse(w, s3err.ErrNoSuchKey, r.URL)
 			return
@@ -326,7 +326,11 @@ func passThroughResponse(proxyResponse *http.Response, w http.ResponseWriter) {
 	for k, v := range proxyResponse.Header {
 		w.Header()[k] = v
 	}
-	w.WriteHeader(proxyResponse.StatusCode)
+	if proxyResponse.Header.Get("Content-Range") != "" && proxyResponse.StatusCode == 200 {
+		w.WriteHeader(http.StatusPartialContent)
+	} else {
+		w.WriteHeader(proxyResponse.StatusCode)
+	}
 	io.Copy(w, proxyResponse.Body)
 }
 

weed/s3api/s3api_objects_list_handlers.go

@@ -206,7 +206,6 @@ func (s3a *S3ApiServer) doListFilerEntries(client filer_pb.SeaweedFilerClient, d
 		isTruncated = isTruncated || subIsTruncated
 		maxKeys -= subCounter
 		nextMarker = subDir + "/" + subNextMarker
-		counter += subCounter
 		// finished processing this sub directory
 		marker = subDir
 	}

weed/sequence/snowflake_sequencer.go

@@ -0,0 +1,46 @@
+package sequence
+
+import (
+	"fmt"
+	"hash/fnv"
+
+	"github.com/bwmarrin/snowflake"
+	"github.com/chrislusf/seaweedfs/weed/glog"
+)
+
+// a simple snowflake Sequencer
+type SnowflakeSequencer struct {
+	node *snowflake.Node
+}
+
+func NewSnowflakeSequencer(nodeid string) (*SnowflakeSequencer, error) {
+	nodeid_hash := hash(nodeid) & 0x3ff
+	glog.V(0).Infof("use snowflake seq id generator, nodeid:%s hex_of_nodeid: %x", nodeid, nodeid_hash)
+	node, err := snowflake.NewNode(int64(nodeid_hash))
+	if err != nil {
+		fmt.Println(err)
+		return nil, err
+	}
+
+	sequencer := &SnowflakeSequencer{node: node}
+	return sequencer, nil
+}
+
+func hash(s string) uint32 {
+	h := fnv.New32a()
+	h.Write([]byte(s))
+	return h.Sum32()
+}
+
+func (m *SnowflakeSequencer) NextFileId(count uint64) uint64 {
+	return uint64(m.node.Generate().Int64())
+}
+
+// ignore setmax as we are snowflake
+func (m *SnowflakeSequencer) SetMax(seenValue uint64) {
+}
+
+// return a new id as no Peek is stored
+func (m *SnowflakeSequencer) Peek() uint64 {
+	return uint64(m.node.Generate().Int64())
+}

weed/server/common.go

@@ -234,12 +234,12 @@ func adjustHeaderContentDisposition(w http.ResponseWriter, r *http.Request, file
 	}
 }
 
-func processRangeRequest(r *http.Request, w http.ResponseWriter, totalSize int64, mimeType string, writeFn func(writer io.Writer, offset int64, size int64, httpStatusCode int) error) {
+func processRangeRequest(r *http.Request, w http.ResponseWriter, totalSize int64, mimeType string, writeFn func(writer io.Writer, offset int64, size int64) error) {
 	rangeReq := r.Header.Get("Range")
 
 	if rangeReq == "" {
 		w.Header().Set("Content-Length", strconv.FormatInt(totalSize, 10))
-		if err := writeFn(w, 0, totalSize, 0); err != nil {
+		if err := writeFn(w, 0, totalSize); err != nil {
 			http.Error(w, err.Error(), http.StatusInternalServerError)
 			return
 		}
@@ -279,7 +279,7 @@ func processRangeRequest(r *http.Request, w http.ResponseWriter, totalSize int64
 		w.Header().Set("Content-Length", strconv.FormatInt(ra.length, 10))
 		w.Header().Set("Content-Range", ra.contentRange(totalSize))
 
-		err = writeFn(w, ra.start, ra.length, http.StatusPartialContent)
+		err = writeFn(w, ra.start, ra.length)
 		if err != nil {
 			http.Error(w, err.Error(), http.StatusInternalServerError)
 			return
@@ -307,7 +307,7 @@ func processRangeRequest(r *http.Request, w http.ResponseWriter, totalSize int64
 				pw.CloseWithError(e)
 				return
 			}
-			if e = writeFn(part, ra.start, ra.length, 0); e != nil {
+			if e = writeFn(part, ra.start, ra.length); e != nil {
 				pw.CloseWithError(e)
 				return
 			}

weed/server/filer_grpc_server_rename.go

@@ -33,8 +33,7 @@ func (fs *FilerServer) AtomicRenameEntry(ctx context.Context, req *filer_pb.Atom
 		return nil, fmt.Errorf("%s/%s not found: %v", req.OldDirectory, req.OldName, err)
 	}
 
-	var events MoveEvents
-	moveErr := fs.moveEntry(ctx, oldParent, oldEntry, newParent, req.NewName, &events)
+	moveErr := fs.moveEntry(ctx, oldParent, oldEntry, newParent, req.NewName)
 	if moveErr != nil {
 		fs.filer.RollbackTransaction(ctx)
 		return nil, fmt.Errorf("%s/%s move error: %v", req.OldDirectory, req.OldName, moveErr)
@@ -48,11 +47,11 @@ func (fs *FilerServer) AtomicRenameEntry(ctx context.Context, req *filer_pb.Atom
 	return &filer_pb.AtomicRenameEntryResponse{}, nil
 }
 
-func (fs *FilerServer) moveEntry(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string, events *MoveEvents) error {
+func (fs *FilerServer) moveEntry(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string) error {
 
-	if err := fs.moveSelfEntry(ctx, oldParent, entry, newParent, newName, events, func() error {
+	if err := fs.moveSelfEntry(ctx, oldParent, entry, newParent, newName, func() error {
 		if entry.IsDirectory() {
-			if err := fs.moveFolderSubEntries(ctx, oldParent, entry, newParent, newName, events); err != nil {
+			if err := fs.moveFolderSubEntries(ctx, oldParent, entry, newParent, newName); err != nil {
 				return err
 			}
 		}
@@ -64,7 +63,7 @@ func (fs *FilerServer) moveEntry(ctx context.Context, oldParent util.FullPath, e
 	return nil
 }
 
-func (fs *FilerServer) moveFolderSubEntries(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string, events *MoveEvents) error {
+func (fs *FilerServer) moveFolderSubEntries(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string) error {
 
 	currentDirPath := oldParent.Child(entry.Name())
 	newDirPath := newParent.Child(newName)
@@ -85,7 +84,7 @@ func (fs *FilerServer) moveFolderSubEntries(ctx context.Context, oldParent util.
 		for _, item := range entries {
 			lastFileName = item.Name()
 			// println("processing", lastFileName)
-			err := fs.moveEntry(ctx, currentDirPath, item, newDirPath, item.Name(), events)
+			err := fs.moveEntry(ctx, currentDirPath, item, newDirPath, item.Name())
 			if err != nil {
 				return err
 			}
@@ -97,8 +96,7 @@ func (fs *FilerServer) moveFolderSubEntries(ctx context.Context, oldParent util.
 	return nil
 }
 
-func (fs *FilerServer) moveSelfEntry(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string, events *MoveEvents,
-	moveFolderSubEntries func() error) error {
+func (fs *FilerServer) moveSelfEntry(ctx context.Context, oldParent util.FullPath, entry *filer.Entry, newParent util.FullPath, newName string, moveFolderSubEntries func() error) error {
 
 	oldPath, newPath := oldParent.Child(entry.Name()), newParent.Child(newName)
 
@@ -122,8 +120,6 @@ func (fs *FilerServer) moveSelfEntry(ctx context.Context, oldParent util.FullPat
 		return createErr
 	}
 
-	events.newEntries = append(events.newEntries, newEntry)
-
 	if moveFolderSubEntries != nil {
 		if moveChildrenErr := moveFolderSubEntries(); moveChildrenErr != nil {
 			return moveChildrenErr
@@ -136,13 +132,6 @@ func (fs *FilerServer) moveSelfEntry(ctx context.Context, oldParent util.FullPat
 		return deleteErr
 	}
 
-	events.oldEntries = append(events.oldEntries, entry)
-
 	return nil
 
 }
-
-type MoveEvents struct {
-	oldEntries []*filer.Entry
-	newEntries []*filer.Entry
-}

weed/server/filer_server.go

@@ -45,22 +45,23 @@ import (
 )
 
 type FilerOption struct {
-	Masters            []string
-	Collection         string
-	DefaultReplication string
-	DisableDirListing  bool
-	MaxMB              int
-	DirListingLimit    int
-	DataCenter         string
-	Rack               string
-	DefaultLevelDbDir  string
-	DisableHttp        bool
-	Host               string
-	Port               uint32
-	recursiveDelete    bool
-	Cipher             bool
-	SaveToFilerLimit   int
-	Filers             []string
+	Masters               []string
+	Collection            string
+	DefaultReplication    string
+	DisableDirListing     bool
+	MaxMB                 int
+	DirListingLimit       int
+	DataCenter            string
+	Rack                  string
+	DefaultLevelDbDir     string
+	DisableHttp           bool
+	Host                  string
+	Port                  uint32
+	recursiveDelete       bool
+	Cipher                bool
+	SaveToFilerLimit      int64
+	Filers                []string
+	ConcurrentUploadLimit int64
 }
 
 type FilerServer struct {
@@ -79,14 +80,18 @@ type FilerServer struct {
 
 	brokers     map[string]map[string]bool
 	brokersLock sync.Mutex
+
+	inFlightDataSize      int64
+	inFlightDataLimitCond *sync.Cond
 }
 
 func NewFilerServer(defaultMux, readonlyMux *http.ServeMux, option *FilerOption) (fs *FilerServer, err error) {
 
 	fs = &FilerServer{
-		option:         option,
-		grpcDialOption: security.LoadClientTLS(util.GetViper(), "grpc.filer"),
-		brokers:        make(map[string]map[string]bool),
+		option:                option,
+		grpcDialOption:        security.LoadClientTLS(util.GetViper(), "grpc.filer"),
+		brokers:               make(map[string]map[string]bool),
+		inFlightDataLimitCond: sync.NewCond(new(sync.Mutex)),
 	}
 	fs.listenersCond = sync.NewCond(&fs.listenersLock)
 

weed/server/filer_server_handlers.go

@@ -4,6 +4,7 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/util"
 	"net/http"
 	"strings"
+	"sync/atomic"
 	"time"
 
 	"github.com/chrislusf/seaweedfs/weed/stats"
@@ -47,18 +48,34 @@ func (fs *FilerServer) filerHandler(w http.ResponseWriter, r *http.Request) {
 			fs.DeleteHandler(w, r)
 		}
 		stats.FilerRequestHistogram.WithLabelValues("delete").Observe(time.Since(start).Seconds())
-	case "PUT":
-		stats.FilerRequestCounter.WithLabelValues("put").Inc()
-		if _, ok := r.URL.Query()["tagging"]; ok {
-			fs.PutTaggingHandler(w, r)
-		} else {
-			fs.PostHandler(w, r)
+	case "POST", "PUT":
+
+		// wait until in flight data is less than the limit
+		contentLength := getContentLength(r)
+		fs.inFlightDataLimitCond.L.Lock()
+		for atomic.LoadInt64(&fs.inFlightDataSize) > fs.option.ConcurrentUploadLimit {
+			fs.inFlightDataLimitCond.Wait()
+		}
+		atomic.AddInt64(&fs.inFlightDataSize, contentLength)
+		fs.inFlightDataLimitCond.L.Unlock()
+		defer func() {
+			atomic.AddInt64(&fs.inFlightDataSize, -contentLength)
+			fs.inFlightDataLimitCond.Signal()
+		}()
+
+		if r.Method == "PUT" {
+			stats.FilerRequestCounter.WithLabelValues("put").Inc()
+			if _, ok := r.URL.Query()["tagging"]; ok {
+				fs.PutTaggingHandler(w, r)
+			} else {
+				fs.PostHandler(w, r, contentLength)
+			}
+			stats.FilerRequestHistogram.WithLabelValues("put").Observe(time.Since(start).Seconds())
+		} else { // method == "POST"
+			stats.FilerRequestCounter.WithLabelValues("post").Inc()
+			fs.PostHandler(w, r, contentLength)
+			stats.FilerRequestHistogram.WithLabelValues("post").Observe(time.Since(start).Seconds())
 		}
-		stats.FilerRequestHistogram.WithLabelValues("put").Observe(time.Since(start).Seconds())
-	case "POST":
-		stats.FilerRequestCounter.WithLabelValues("post").Inc()
-		fs.PostHandler(w, r)
-		stats.FilerRequestHistogram.WithLabelValues("post").Observe(time.Since(start).Seconds())
 	case "OPTIONS":
 		stats.FilerRequestCounter.WithLabelValues("options").Inc()
 		OptionsHandler(w, r, false)

weed/server/filer_server_handlers_read.go

@@ -131,6 +131,9 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request,
 
 	if r.Method == "HEAD" {
 		w.Header().Set("Content-Length", strconv.FormatInt(totalSize, 10))
+		processRangeRequest(r, w, totalSize, mimeType, func(writer io.Writer, offset int64, size int64) error {
+			return filer.StreamContent(fs.filer.MasterClient, writer, entry.Chunks, offset, size, true)
+		})
 		return
 	}
 
@@ -150,10 +153,7 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request,
 		}
 	}
 
-	processRangeRequest(r, w, totalSize, mimeType, func(writer io.Writer, offset int64, size int64, httpStatusCode int) error {
-		if httpStatusCode != 0 {
-			w.WriteHeader(httpStatusCode)
-		}
+	processRangeRequest(r, w, totalSize, mimeType, func(writer io.Writer, offset int64, size int64) error {
 		if offset+size <= int64(len(entry.Content)) {
 			_, err := writer.Write(entry.Content[offset : offset+size])
 			if err != nil {
@@ -161,7 +161,7 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request,
 			}
 			return err
 		}
-		return filer.StreamContent(fs.filer.MasterClient, writer, entry.Chunks, offset, size)
+		return filer.StreamContent(fs.filer.MasterClient, writer, entry.Chunks, offset, size, false)
 	})
 
 }

weed/server/filer_server_handlers_write.go

@@ -52,7 +52,7 @@ func (fs *FilerServer) assignNewFileInfo(so *operation.StorageOption) (fileId, u
 	return
 }
 
-func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) {
+func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request, contentLength int64) {
 
 	ctx := context.Background()
 
@@ -66,7 +66,7 @@ func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) {
 		query.Get("rack"),
 	)
 
-	fs.autoChunk(ctx, w, r, so)
+	fs.autoChunk(ctx, w, r, contentLength, so)
 	util.CloseRequest(r)
 
 }

weed/server/filer_server_handlers_write_autochunk.go

@@ -2,11 +2,8 @@ package weed_server
 
 import (
 	"context"
-	"crypto/md5"
 	"fmt"
-	"hash"
 	"io"
-	"io/ioutil"
 	"net/http"
 	"os"
 	"path"
@@ -19,13 +16,12 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/operation"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
 	xhttp "github.com/chrislusf/seaweedfs/weed/s3api/http"
-	"github.com/chrislusf/seaweedfs/weed/security"
 	"github.com/chrislusf/seaweedfs/weed/stats"
 	"github.com/chrislusf/seaweedfs/weed/storage/needle"
 	"github.com/chrislusf/seaweedfs/weed/util"
 )
 
-func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, so *operation.StorageOption) {
+func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, contentLength int64, so *operation.StorageOption) {
 
 	// autoChunking can be set at the command-line level or as a query param. Query param overrides command-line
 	query := r.URL.Query()
@@ -51,14 +47,16 @@ func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *
 		if r.Header.Get("Content-Type") == "" && strings.HasSuffix(r.URL.Path, "/") {
 			reply, err = fs.mkdir(ctx, w, r)
 		} else {
-			reply, md5bytes, err = fs.doPostAutoChunk(ctx, w, r, chunkSize, so)
+			reply, md5bytes, err = fs.doPostAutoChunk(ctx, w, r, chunkSize, contentLength, so)
 		}
 	} else {
-		reply, md5bytes, err = fs.doPutAutoChunk(ctx, w, r, chunkSize, so)
+		reply, md5bytes, err = fs.doPutAutoChunk(ctx, w, r, chunkSize, contentLength, so)
 	}
 	if err != nil {
 		if strings.HasPrefix(err.Error(), "read input:") {
 			writeJsonError(w, r, 499, err)
+		} else if strings.HasSuffix(err.Error(), "is a file") {
+			writeJsonError(w, r, http.StatusConflict, err)
 		} else {
 			writeJsonError(w, r, http.StatusInternalServerError, err)
 		}
@@ -70,7 +68,7 @@ func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *
 	}
 }
 
-func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
+func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, contentLength int64, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
 
 	multipartReader, multipartReaderErr := r.MultipartReader()
 	if multipartReaderErr != nil {
@@ -91,7 +89,7 @@ func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWrite
 		contentType = ""
 	}
 
-	fileChunks, md5Hash, chunkOffset, err, smallContent := fs.uploadReaderToChunks(w, r, part1, chunkSize, fileName, contentType, so)
+	fileChunks, md5Hash, chunkOffset, err, smallContent := fs.uploadReaderToChunks(w, r, part1, chunkSize, fileName, contentType, contentLength, so)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -102,7 +100,7 @@ func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWrite
 	return
 }
 
-func (fs *FilerServer) doPutAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
+func (fs *FilerServer) doPutAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, contentLength int64, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
 
 	fileName := path.Base(r.URL.Path)
 	contentType := r.Header.Get("Content-Type")
@@ -110,7 +108,7 @@ func (fs *FilerServer) doPutAutoChunk(ctx context.Context, w http.ResponseWriter
 		contentType = ""
 	}
 
-	fileChunks, md5Hash, chunkOffset, err, smallContent := fs.uploadReaderToChunks(w, r, r.Body, chunkSize, fileName, contentType, so)
+	fileChunks, md5Hash, chunkOffset, err, smallContent := fs.uploadReaderToChunks(w, r, r.Body, chunkSize, fileName, contentType, contentLength, so)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -212,7 +210,7 @@ func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileNa
 		entry.Extended = make(map[string][]byte)
 	}
 
-	fs.saveAmzMetaData(r, entry)
+	SaveAmzMetaData(r, entry.Extended, false)
 
 	for k, v := range r.Header {
 		if len(v) > 0 && strings.HasPrefix(k, needle.PairNamePrefix) {
@@ -229,92 +227,6 @@ func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileNa
 	return filerResult, replyerr
 }
 
-func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Request, reader io.Reader, chunkSize int32, fileName, contentType string, so *operation.StorageOption) ([]*filer_pb.FileChunk, hash.Hash, int64, error, []byte) {
-	var fileChunks []*filer_pb.FileChunk
-
-	md5Hash := md5.New()
-	var partReader = ioutil.NopCloser(io.TeeReader(reader, md5Hash))
-
-	chunkOffset := int64(0)
-	var smallContent []byte
-
-	for {
-		limitedReader := io.LimitReader(partReader, int64(chunkSize))
-
-		data, err := ioutil.ReadAll(limitedReader)
-		if err != nil {
-			return nil, nil, 0, err, nil
-		}
-		if chunkOffset == 0 && !isAppend(r) {
-			if len(data) < fs.option.SaveToFilerLimit || strings.HasPrefix(r.URL.Path, filer.DirectoryEtcRoot) && len(data) < 4*1024 {
-				smallContent = data
-				chunkOffset += int64(len(data))
-				break
-			}
-		}
-		dataReader := util.NewBytesReader(data)
-
-		// retry to assign a different file id
-		var fileId, urlLocation string
-		var auth security.EncodedJwt
-		var assignErr, uploadErr error
-		var uploadResult *operation.UploadResult
-		for i := 0; i < 3; i++ {
-			// assign one file id for one chunk
-			fileId, urlLocation, auth, assignErr = fs.assignNewFileInfo(so)
-			if assignErr != nil {
-				return nil, nil, 0, assignErr, nil
-			}
-
-			// upload the chunk to the volume server
-			uploadResult, uploadErr, _ = fs.doUpload(urlLocation, w, r, dataReader, fileName, contentType, nil, auth)
-			if uploadErr != nil {
-				time.Sleep(251 * time.Millisecond)
-				continue
-			}
-			break
-		}
-		if uploadErr != nil {
-			return nil, nil, 0, uploadErr, nil
-		}
-
-		// if last chunk exhausted the reader exactly at the border
-		if uploadResult.Size == 0 {
-			break
-		}
-
-		// Save to chunk manifest structure
-		fileChunks = append(fileChunks, uploadResult.ToPbFileChunk(fileId, chunkOffset))
-
-		glog.V(4).Infof("uploaded %s chunk %d to %s [%d,%d)", fileName, len(fileChunks), fileId, chunkOffset, chunkOffset+int64(uploadResult.Size))
-
-		// reset variables for the next chunk
-		chunkOffset = chunkOffset + int64(uploadResult.Size)
-
-		// if last chunk was not at full chunk size, but already exhausted the reader
-		if int64(uploadResult.Size) < int64(chunkSize) {
-			break
-		}
-	}
-
-	return fileChunks, md5Hash, chunkOffset, nil, smallContent
-}
-
-func (fs *FilerServer) doUpload(urlLocation string, w http.ResponseWriter, r *http.Request, limitedReader io.Reader, fileName string, contentType string, pairMap map[string]string, auth security.EncodedJwt) (*operation.UploadResult, error, []byte) {
-
-	stats.FilerRequestCounter.WithLabelValues("chunkUpload").Inc()
-	start := time.Now()
-	defer func() {
-		stats.FilerRequestHistogram.WithLabelValues("chunkUpload").Observe(time.Since(start).Seconds())
-	}()
-
-	uploadResult, err, data := operation.Upload(urlLocation, fileName, fs.option.Cipher, limitedReader, false, contentType, pairMap, auth)
-	if uploadResult != nil && uploadResult.RetryCount > 0 {
-		stats.FilerRequestCounter.WithLabelValues("chunkUploadRetry").Add(float64(uploadResult.RetryCount))
-	}
-	return uploadResult, err, data
-}
-
 func (fs *FilerServer) saveAsChunk(so *operation.StorageOption) filer.SaveDataAsChunkFunctionType {
 
 	return func(reader io.Reader, name string, offset int64) (*filer_pb.FileChunk, string, string, error) {
@@ -383,17 +295,24 @@ func (fs *FilerServer) mkdir(ctx context.Context, w http.ResponseWriter, r *http
 	return filerResult, replyerr
 }
 
-func (fs *FilerServer) saveAmzMetaData(r *http.Request, entry *filer.Entry) {
+func SaveAmzMetaData(r *http.Request, existing map[string][]byte, isReplace bool) (metadata map[string][]byte) {
+
+	metadata = make(map[string][]byte)
+	if !isReplace {
+		for k, v := range existing {
+			metadata[k] = v
+		}
+	}
 
 	if sc := r.Header.Get(xhttp.AmzStorageClass); sc != "" {
-		entry.Extended[xhttp.AmzStorageClass] = []byte(sc)
+		metadata[xhttp.AmzStorageClass] = []byte(sc)
 	}
 
 	if tags := r.Header.Get(xhttp.AmzObjectTagging); tags != "" {
 		for _, v := range strings.Split(tags, "&") {
 			tag := strings.Split(v, "=")
 			if len(tag) == 2 {
-				entry.Extended[xhttp.AmzObjectTagging+"-"+tag[0]] = []byte(tag[1])
+				metadata[xhttp.AmzObjectTagging+"-"+tag[0]] = []byte(tag[1])
 			}
 		}
 	}
@@ -401,8 +320,11 @@ func (fs *FilerServer) saveAmzMetaData(r *http.Request, entry *filer.Entry) {
 	for header, values := range r.Header {
 		if strings.HasPrefix(header, xhttp.AmzUserMetaPrefix) {
 			for _, value := range values {
-				entry.Extended[header] = []byte(value)
+				metadata[header] = []byte(value)
 			}
 		}
 	}
+
+	return
+
 }

weed/server/filer_server_handlers_write_upload.go

@@ -0,0 +1,175 @@
+package weed_server
+
+import (
+	"crypto/md5"
+	"hash"
+	"io"
+	"io/ioutil"
+	"net/http"
+	"runtime"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/chrislusf/seaweedfs/weed/filer"
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/operation"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/security"
+	"github.com/chrislusf/seaweedfs/weed/stats"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+var (
+	limitedUploadProcessor = util.NewLimitedOutOfOrderProcessor(int32(runtime.NumCPU()))
+)
+
+func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Request, reader io.Reader, chunkSize int32, fileName, contentType string, contentLength int64, so *operation.StorageOption) (fileChunks []*filer_pb.FileChunk, md5Hash hash.Hash, dataSize int64, err error, smallContent []byte) {
+
+	md5Hash = md5.New()
+	var partReader = ioutil.NopCloser(io.TeeReader(reader, md5Hash))
+
+	// save small content directly
+	if !isAppend(r) && ((0 < contentLength && contentLength < fs.option.SaveToFilerLimit) || strings.HasPrefix(r.URL.Path, filer.DirectoryEtcRoot) && contentLength < 4*1024) {
+		smallContent, err = ioutil.ReadAll(partReader)
+		dataSize = int64(len(smallContent))
+		return
+	}
+
+	resultsChan := make(chan *ChunkCreationResult, 2)
+
+	var waitForAllData sync.WaitGroup
+	waitForAllData.Add(1)
+	go func() {
+		// process upload results
+		defer waitForAllData.Done()
+		for result := range resultsChan {
+			if result.err != nil {
+				err = result.err
+				continue
+			}
+
+			// Save to chunk manifest structure
+			fileChunks = append(fileChunks, result.chunk)
+		}
+	}()
+
+	var lock sync.Mutex
+	readOffset := int64(0)
+	var wg sync.WaitGroup
+
+	for err == nil {
+
+		wg.Add(1)
+		request := func() {
+			defer wg.Done()
+
+			var localOffset int64
+			// read from the input
+			lock.Lock()
+			localOffset = readOffset
+			limitedReader := io.LimitReader(partReader, int64(chunkSize))
+			data, readErr := ioutil.ReadAll(limitedReader)
+			readOffset += int64(len(data))
+			lock.Unlock()
+			// handle read errors
+			if readErr != nil {
+				if err == nil {
+					err = readErr
+				}
+				if readErr != io.EOF {
+					resultsChan <- &ChunkCreationResult{
+						err: readErr,
+					}
+				}
+				return
+			}
+			if len(data) == 0 {
+				readErr = io.EOF
+				if err == nil {
+					err = readErr
+				}
+				return
+			}
+
+			// upload
+			dataReader := util.NewBytesReader(data)
+			fileId, uploadResult, uploadErr := fs.doCreateChunk(w, r, so, dataReader, fileName, contentType)
+			if uploadErr != nil {
+				if err == nil {
+					err = uploadErr
+				}
+				resultsChan <- &ChunkCreationResult{
+					err: uploadErr,
+				}
+				return
+			}
+
+			glog.V(4).Infof("uploaded %s to %s [%d,%d)", fileName, fileId, localOffset, localOffset+int64(uploadResult.Size))
+
+			// send back uploaded file chunk
+			resultsChan <- &ChunkCreationResult{
+				chunk: uploadResult.ToPbFileChunk(fileId, localOffset),
+			}
+
+		}
+		limitedUploadProcessor.Execute(request)
+	}
+
+	go func() {
+		wg.Wait()
+		close(resultsChan)
+	}()
+
+	waitForAllData.Wait()
+
+	if err == io.EOF {
+		err = nil
+	}
+
+	return fileChunks, md5Hash, readOffset, err, nil
+}
+
+type ChunkCreationResult struct {
+	chunk *filer_pb.FileChunk
+	err   error
+}
+
+func (fs *FilerServer) doCreateChunk(w http.ResponseWriter, r *http.Request, so *operation.StorageOption, dataReader *util.BytesReader, fileName string, contentType string) (string, *operation.UploadResult, error) {
+	// retry to assign a different file id
+	var fileId, urlLocation string
+	var auth security.EncodedJwt
+	var assignErr, uploadErr error
+	var uploadResult *operation.UploadResult
+	for i := 0; i < 3; i++ {
+		// assign one file id for one chunk
+		fileId, urlLocation, auth, assignErr = fs.assignNewFileInfo(so)
+		if assignErr != nil {
+			return "", nil, assignErr
+		}
+
+		// upload the chunk to the volume server
+		uploadResult, uploadErr, _ = fs.doUpload(urlLocation, w, r, dataReader, fileName, contentType, nil, auth)
+		if uploadErr != nil {
+			time.Sleep(251 * time.Millisecond)
+			continue
+		}
+		break
+	}
+	return fileId, uploadResult, uploadErr
+}
+
+func (fs *FilerServer) doUpload(urlLocation string, w http.ResponseWriter, r *http.Request, limitedReader io.Reader, fileName string, contentType string, pairMap map[string]string, auth security.EncodedJwt) (*operation.UploadResult, error, []byte) {
+
+	stats.FilerRequestCounter.WithLabelValues("chunkUpload").Inc()
+	start := time.Now()
+	defer func() {
+		stats.FilerRequestHistogram.WithLabelValues("chunkUpload").Observe(time.Since(start).Seconds())
+	}()
+
+	uploadResult, err, data := operation.Upload(urlLocation, fileName, fs.option.Cipher, limitedReader, false, contentType, pairMap, auth)
+	if uploadResult != nil && uploadResult.RetryCount > 0 {
+		stats.FilerRequestCounter.WithLabelValues("chunkUploadRetry").Add(float64(uploadResult.RetryCount))
+	}
+	return uploadResult, err, data
+}

weed/server/master_grpc_server.go

@@ -80,10 +80,14 @@ func (ms *MasterServer) SendHeartbeat(stream master_pb.Seaweed_SendHeartbeatServ
 		dn.AdjustMaxVolumeCounts(heartbeat.MaxVolumeCounts)
 
 		glog.V(4).Infof("master received heartbeat %s", heartbeat.String())
+		var dataCenter string
+		if dc := dn.GetDataCenter(); dc != nil {
+			dataCenter = string(dc.Id())
+		}
 		message := &master_pb.VolumeLocation{
 			Url:        dn.Url(),
 			PublicUrl:  dn.PublicUrl,
-			DataCenter: string(dn.GetDataCenter().Id()),
+			DataCenter: dataCenter,
 		}
 		if len(heartbeat.NewVolumes) > 0 || len(heartbeat.DeletedVolumes) > 0 {
 			// process delta volume ids if exists for fast volume id updates

weed/server/master_server.go

@@ -277,6 +277,13 @@ func (ms *MasterServer) createSequencer(option *MasterOption) sequence.Sequencer
 			glog.Error(err)
 			seq = nil
 		}
+	case "snowflake":
+		var err error
+		seq, err = sequence.NewSnowflakeSequencer(fmt.Sprintf("%s:%d", option.Host, option.Port))
+		if err != nil {
+			glog.Error(err)
+			seq = nil
+		}
 	default:
 		seq = sequence.NewMemorySequencer()
 	}

weed/server/volume_grpc_read_write.go

@@ -0,0 +1,38 @@
+package weed_server
+
+import (
+	"context"
+	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/pb/volume_server_pb"
+	"github.com/chrislusf/seaweedfs/weed/storage/needle"
+	"github.com/chrislusf/seaweedfs/weed/storage/types"
+)
+
+func (vs *VolumeServer) ReadNeedleBlob(ctx context.Context, req *volume_server_pb.ReadNeedleBlobRequest) (resp *volume_server_pb.ReadNeedleBlobResponse, err error) {
+	resp = &volume_server_pb.ReadNeedleBlobResponse{}
+	v := vs.store.GetVolume(needle.VolumeId(req.VolumeId))
+	if v == nil {
+		return nil, fmt.Errorf("not found volume id %d", req.VolumeId)
+	}
+
+	resp.NeedleBlob, err = v.ReadNeedleBlob(req.Offset, types.Size(req.Size))
+	if err != nil {
+		return nil, fmt.Errorf("read needle blob offset %d size %d: %v", req.Offset, req.Size, err)
+	}
+
+	return resp, nil
+}
+
+func (vs *VolumeServer) WriteNeedleBlob(ctx context.Context, req *volume_server_pb.WriteNeedleBlobRequest) (resp *volume_server_pb.WriteNeedleBlobResponse, err error) {
+	resp = &volume_server_pb.WriteNeedleBlobResponse{}
+	v := vs.store.GetVolume(needle.VolumeId(req.VolumeId))
+	if v == nil {
+		return nil, fmt.Errorf("not found volume id %d", req.VolumeId)
+	}
+
+	if err = v.WriteNeedleBlob(types.NeedleId(req.NeedleId), req.NeedleBlob, types.Size(req.Size)); err != nil {
+		return nil, fmt.Errorf("write blob needle %d size %d: %v", req.NeedleId, req.Size, err)
+	}
+
+	return resp, nil
+}

weed/server/volume_server.go

@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"github.com/chrislusf/seaweedfs/weed/storage/types"
 	"net/http"
+	"sync"
 
 	"google.golang.org/grpc"
 
@@ -34,6 +35,10 @@ type VolumeServer struct {
 	fileSizeLimitBytes      int64
 	isHeartbeating          bool
 	stopChan                chan bool
+
+	inFlightDataSize      int64
+	inFlightDataLimitCond *sync.Cond
+	concurrentUploadLimit int64
 }
 
 func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
@@ -48,6 +53,7 @@ func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
 	readRedirect bool,
 	compactionMBPerSecond int,
 	fileSizeLimitMB int,
+	concurrentUploadLimit int64,
 ) *VolumeServer {
 
 	v := util.GetViper()
@@ -72,6 +78,8 @@ func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
 		fileSizeLimitBytes:      int64(fileSizeLimitMB) * 1024 * 1024,
 		isHeartbeating:          true,
 		stopChan:                make(chan bool),
+		inFlightDataLimitCond:   sync.NewCond(new(sync.Mutex)),
+		concurrentUploadLimit:   concurrentUploadLimit,
 	}
 	vs.SeedMasterNodes = masterNodes
 

weed/server/volume_server_handlers.go

@@ -2,7 +2,9 @@ package weed_server
 
 import (
 	"net/http"
+	"strconv"
 	"strings"
+	"sync/atomic"
 
 	"github.com/chrislusf/seaweedfs/weed/util"
 
@@ -40,8 +42,24 @@ func (vs *VolumeServer) privateStoreHandler(w http.ResponseWriter, r *http.Reque
 		stats.DeleteRequest()
 		vs.guard.WhiteList(vs.DeleteHandler)(w, r)
 	case "PUT", "POST":
+
+		// wait until in flight data is less than the limit
+		contentLength := getContentLength(r)
+		vs.inFlightDataLimitCond.L.Lock()
+		for atomic.LoadInt64(&vs.inFlightDataSize) > vs.concurrentUploadLimit {
+			vs.inFlightDataLimitCond.Wait()
+		}
+		atomic.AddInt64(&vs.inFlightDataSize, contentLength)
+		vs.inFlightDataLimitCond.L.Unlock()
+		defer func() {
+			atomic.AddInt64(&vs.inFlightDataSize, -contentLength)
+			vs.inFlightDataLimitCond.Signal()
+		}()
+
+		// processs uploads
 		stats.WriteRequest()
 		vs.guard.WhiteList(vs.PostHandler)(w, r)
+
 	case "OPTIONS":
 		stats.ReadRequest()
 		w.Header().Add("Access-Control-Allow-Methods", "PUT, POST, GET, DELETE, OPTIONS")
@@ -49,6 +67,18 @@ func (vs *VolumeServer) privateStoreHandler(w http.ResponseWriter, r *http.Reque
 	}
 }
 
+func getContentLength(r *http.Request) int64 {
+	contentLength := r.Header.Get("Content-Length")
+	if contentLength != "" {
+		length, err := strconv.ParseInt(contentLength, 10, 64)
+		if err != nil {
+			return 0
+		}
+		return length
+	}
+	return 0
+}
+
 func (vs *VolumeServer) publicReadOnlyHandler(w http.ResponseWriter, r *http.Request) {
 	w.Header().Set("Server", "SeaweedFS Volume "+util.VERSION)
 	if r.Header.Get("Origin") != "" {

weed/server/volume_server_handlers_read.go

@@ -27,7 +27,7 @@ var fileNameEscaper = strings.NewReplacer(`\`, `\\`, `"`, `\"`)
 
 func (vs *VolumeServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request) {
 
-	// println(r.Method + " " + r.URL.Path)
+	glog.V(9).Info(r.Method + " " + r.URL.Path + " " + r.Header.Get("Range"))
 
 	stats.VolumeServerRequestCounter.WithLabelValues("get").Inc()
 	start := time.Now()
@@ -261,13 +261,10 @@ func writeResponseContent(filename, mimeType string, rs io.ReadSeeker, w http.Re
 		return nil
 	}
 
-	processRangeRequest(r, w, totalSize, mimeType, func(writer io.Writer, offset int64, size int64, httpStatusCode int) error {
+	processRangeRequest(r, w, totalSize, mimeType, func(writer io.Writer, offset int64, size int64) error {
 		if _, e = rs.Seek(offset, 0); e != nil {
 			return e
 		}
-		if httpStatusCode != 0 {
-			w.WriteHeader(httpStatusCode)
-		}
 		_, e = io.CopyN(writer, rs, size)
 		return e
 	})

weed/shell/command_fs_cat.go

@@ -52,7 +52,7 @@ func (c *commandFsCat) Do(args []string, commandEnv *CommandEnv, writer io.Write
 			return err
 		}
 
-		return filer.StreamContent(commandEnv.MasterClient, writer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64)
+		return filer.StreamContent(commandEnv.MasterClient, writer, respLookupEntry.Entry.Chunks, 0, math.MaxInt64, false)
 
 	})
 

weed/shell/command_fs_mv.go

@@ -36,6 +36,10 @@ func (c *commandFsMv) Help() string {
 
 func (c *commandFsMv) Do(args []string, commandEnv *CommandEnv, writer io.Writer) (err error) {
 
+	if len(args) != 2 {
+		return fmt.Errorf("need to have 2 arguments")
+	}
+
 	sourcePath, err := commandEnv.parseUrl(args[0])
 	if err != nil {
 		return err

weed/shell/command_volume_balance_test.go

@@ -180,4 +180,4 @@ func TestBalance(t *testing.T) {
 		t.Errorf("balance: %v", err)
 	}
 
-}
+}

weed/shell/command_volume_check_disk.go

@@ -0,0 +1,257 @@
+package shell
+
+import (
+	"bytes"
+	"context"
+	"flag"
+	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/operation"
+	"github.com/chrislusf/seaweedfs/weed/pb/volume_server_pb"
+	"github.com/chrislusf/seaweedfs/weed/storage/needle_map"
+	"io"
+	"math"
+	"sort"
+)
+
+func init() {
+	Commands = append(Commands, &commandVolumeCheckDisk{})
+}
+
+type commandVolumeCheckDisk struct {
+	env *CommandEnv
+}
+
+func (c *commandVolumeCheckDisk) Name() string {
+	return "volume.check.disk"
+}
+
+func (c *commandVolumeCheckDisk) Help() string {
+	return `check all replicated volumes to find and fix inconsistencies
+
+	How it works:
+	
+	find all volumes that are replicated
+	  for each volume id, if there are more than 2 replicas, find one pair with the largest 2 in file count.
+      for the pair volume A and B
+        append entries in A and not in B to B
+        append entries in B and not in A to A
+
+`
+}
+
+func (c *commandVolumeCheckDisk) Do(args []string, commandEnv *CommandEnv, writer io.Writer) (err error) {
+
+	if err = commandEnv.confirmIsLocked(); err != nil {
+		return
+	}
+
+	fsckCommand := flag.NewFlagSet(c.Name(), flag.ContinueOnError)
+	slowMode := fsckCommand.Bool("slow", false, "slow mode checks all replicas even file counts are the same")
+	verbose := fsckCommand.Bool("v", false, "verbose mode")
+	applyChanges := fsckCommand.Bool("force", false, "apply the fix")
+	nonRepairThreshold := fsckCommand.Float64("nonRepairThreshold", 0.3, "repair when missing keys is not more than this limit")
+	if err = fsckCommand.Parse(args); err != nil {
+		return nil
+	}
+
+	c.env = commandEnv
+
+	// collect topology information
+	topologyInfo, _, err := collectTopologyInfo(commandEnv)
+	if err != nil {
+		return err
+	}
+	volumeReplicas, _ := collectVolumeReplicaLocations(topologyInfo)
+
+	// pick 1 pairs of volume replica
+	fileCount := func(replica *VolumeReplica) uint64 {
+		return replica.info.FileCount - replica.info.DeleteCount
+	}
+	aDB, bDB := needle_map.NewMemDb(), needle_map.NewMemDb()
+	defer aDB.Close()
+	defer bDB.Close()
+
+	for _, replicas := range volumeReplicas {
+		sort.Slice(replicas, func(i, j int) bool {
+			return fileCount(replicas[i]) > fileCount(replicas[j])
+		})
+		for len(replicas) >= 2 {
+			a, b := replicas[0], replicas[1]
+			if !*slowMode {
+				if fileCount(a) == fileCount(b) {
+					replicas = replicas[1:]
+					continue
+				}
+			}
+			if a.info.ReadOnly || b.info.ReadOnly {
+				fmt.Fprintf(writer, "skipping readonly volume %d on %s and %s\n", a.info.Id, a.location.dataNode.Id, b.location.dataNode.Id)
+				continue
+			}
+
+			// reset index db
+			aDB.Close()
+			bDB.Close()
+			aDB, bDB = needle_map.NewMemDb(), needle_map.NewMemDb()
+
+			// read index db
+			if err := c.readIndexDatabase(aDB, a.info.Collection, a.info.Id, a.location.dataNode.Id, *verbose, writer); err != nil {
+				return err
+			}
+			if err := c.readIndexDatabase(bDB, b.info.Collection, b.info.Id, b.location.dataNode.Id, *verbose, writer); err != nil {
+				return err
+			}
+
+			// find and make up the differnces
+			if err := c.doVolumeCheckDisk(aDB, bDB, a, b, *verbose, writer, *applyChanges, *nonRepairThreshold); err != nil {
+				return err
+			}
+			if err := c.doVolumeCheckDisk(bDB, aDB, b, a, *verbose, writer, *applyChanges, *nonRepairThreshold); err != nil {
+				return err
+			}
+			replicas = replicas[1:]
+		}
+	}
+
+	return nil
+}
+
+func (c *commandVolumeCheckDisk) doVolumeCheckDisk(subtrahend, minuend *needle_map.MemDb, source, target *VolumeReplica, verbose bool, writer io.Writer, applyChanges bool, nonRepairThreshold float64) error {
+
+	// find missing keys
+	// hash join, can be more efficient
+	var missingNeedles []needle_map.NeedleValue
+	var counter int
+	subtrahend.AscendingVisit(func(value needle_map.NeedleValue) error {
+		counter++
+		if _, found := minuend.Get(value.Key); !found {
+			missingNeedles = append(missingNeedles, value)
+		}
+		return nil
+	})
+
+	fmt.Fprintf(writer, "volume %d %s has %d entries, %s missed %d entries\n", source.info.Id, source.location.dataNode.Id, counter, target.location.dataNode.Id, len(missingNeedles))
+
+	if counter == 0 || len(missingNeedles) == 0 {
+		return nil
+	}
+
+	missingNeedlesFraction := float64(len(missingNeedles)) / float64(counter)
+	if missingNeedlesFraction > nonRepairThreshold {
+		return fmt.Errorf(
+			"failed to start repair volume %d, percentage of missing keys is greater than the threshold: %.2f > %.2f",
+			source.info.Id, missingNeedlesFraction, nonRepairThreshold)
+	}
+
+	for _, needleValue := range missingNeedles {
+
+		needleBlob, err := c.readSourceNeedleBlob(source.location.dataNode.Id, source.info.Id, needleValue)
+		if err != nil {
+			return err
+		}
+
+		if !applyChanges {
+			continue
+		}
+
+		if verbose {
+			fmt.Fprintf(writer, "read %d,%x %s => %s \n", source.info.Id, needleValue.Key, source.location.dataNode.Id, target.location.dataNode.Id)
+		}
+
+		if err := c.writeNeedleBlobToTarget(target.location.dataNode.Id, source.info.Id, needleValue, needleBlob); err != nil {
+			return err
+		}
+
+	}
+
+	return nil
+}
+
+func (c *commandVolumeCheckDisk) readSourceNeedleBlob(sourceVolumeServer string, volumeId uint32, needleValue needle_map.NeedleValue) (needleBlob []byte, err error) {
+
+	err = operation.WithVolumeServerClient(sourceVolumeServer, c.env.option.GrpcDialOption, func(client volume_server_pb.VolumeServerClient) error {
+		resp, err := client.ReadNeedleBlob(context.Background(), &volume_server_pb.ReadNeedleBlobRequest{
+			VolumeId: volumeId,
+			NeedleId: uint64(needleValue.Key),
+			Offset:   needleValue.Offset.ToActualOffset(),
+			Size:     int32(needleValue.Size),
+		})
+		if err != nil {
+			return err
+		}
+		needleBlob = resp.NeedleBlob
+		return nil
+	})
+	return
+}
+
+func (c *commandVolumeCheckDisk) writeNeedleBlobToTarget(targetVolumeServer string, volumeId uint32, needleValue needle_map.NeedleValue, needleBlob []byte) error {
+
+	return operation.WithVolumeServerClient(targetVolumeServer, c.env.option.GrpcDialOption, func(client volume_server_pb.VolumeServerClient) error {
+		_, err := client.WriteNeedleBlob(context.Background(), &volume_server_pb.WriteNeedleBlobRequest{
+			VolumeId:   volumeId,
+			NeedleId:   uint64(needleValue.Key),
+			Size:       int32(needleValue.Size),
+			NeedleBlob: needleBlob,
+		})
+		return err
+	})
+
+}
+
+func (c *commandVolumeCheckDisk) readIndexDatabase(db *needle_map.MemDb, collection string, volumeId uint32, volumeServer string, verbose bool, writer io.Writer) error {
+
+	var buf bytes.Buffer
+	if err := c.copyVolumeIndexFile(collection, volumeId, volumeServer, &buf, verbose, writer); err != nil {
+		return err
+	}
+
+	if verbose {
+		fmt.Fprintf(writer, "load collection %s volume %d index size %d from %s ...\n", collection, volumeId, buf.Len(), volumeServer)
+	}
+
+	return db.LoadFromReaderAt(bytes.NewReader(buf.Bytes()))
+
+}
+
+func (c *commandVolumeCheckDisk) copyVolumeIndexFile(collection string, volumeId uint32, volumeServer string, buf *bytes.Buffer, verbose bool, writer io.Writer) error {
+
+	return operation.WithVolumeServerClient(volumeServer, c.env.option.GrpcDialOption, func(volumeServerClient volume_server_pb.VolumeServerClient) error {
+
+		ext := ".idx"
+
+		copyFileClient, err := volumeServerClient.CopyFile(context.Background(), &volume_server_pb.CopyFileRequest{
+			VolumeId:                 volumeId,
+			Ext:                      ".idx",
+			CompactionRevision:       math.MaxUint32,
+			StopOffset:               math.MaxInt64,
+			Collection:               collection,
+			IsEcVolume:               false,
+			IgnoreSourceFileNotFound: false,
+		})
+		if err != nil {
+			return fmt.Errorf("failed to start copying volume %d%s: %v", volumeId, ext, err)
+		}
+
+		err = writeToBuffer(copyFileClient, buf)
+		if err != nil {
+			return fmt.Errorf("failed to copy %d%s from %s: %v", volumeId, ext, volumeServer, err)
+		}
+
+		return nil
+
+	})
+}
+
+func writeToBuffer(client volume_server_pb.VolumeServer_CopyFileClient, buf *bytes.Buffer) error {
+	for {
+		resp, receiveErr := client.Recv()
+		if receiveErr == io.EOF {
+			break
+		}
+		if receiveErr != nil {
+			return fmt.Errorf("receiving: %v", receiveErr)
+		}
+		buf.Write(resp.FileContent)
+	}
+	return nil
+}

weed/shell/command_volume_list_test.go

@@ -67,7 +67,7 @@ func parseOutput(output string) *master_pb.TopologyInfo {
 				maxVolumeCountStr := parts[1][strings.Index(parts[1], "/")+1:]
 				maxVolumeCount, _ := strconv.Atoi(maxVolumeCountStr)
 				disk = &master_pb.DiskInfo{
-					Type: diskType,
+					Type:           diskType,
 					MaxVolumeCount: uint64(maxVolumeCount),
 				}
 				dn.DiskInfos[types.ToDiskType(diskType).String()] = disk

weed/shell/shell_liner.go

@@ -19,6 +19,10 @@ var (
 
 func RunShell(options ShellOptions) {
 
+	sort.Slice(Commands, func(i, j int) bool {
+		return strings.Compare(Commands[i].Name(), Commands[j].Name()) < 0
+	})
+
 	line = liner.NewLiner()
 	defer line.Close()
 
@@ -96,9 +100,6 @@ func printGenericHelp() {
 `
 	fmt.Print(msg)
 
-	sort.Slice(Commands, func(i, j int) bool {
-		return strings.Compare(Commands[i].Name(), Commands[j].Name()) < 0
-	})
 	for _, c := range Commands {
 		helpTexts := strings.SplitN(c.Help(), "\n", 2)
 		fmt.Printf("  %-30s\t# %s \n", c.Name(), helpTexts[0])
@@ -114,10 +115,6 @@ func printHelp(cmds []string) {
 	} else {
 		cmd := strings.ToLower(args[0])
 
-		sort.Slice(Commands, func(i, j int) bool {
-			return strings.Compare(Commands[i].Name(), Commands[j].Name()) < 0
-		})
-
 		for _, c := range Commands {
 			if c.Name() == cmd {
 				fmt.Printf("  %s\t# %s\n", c.Name(), c.Help())

weed/storage/needle/needle_read_write.go

@@ -3,13 +3,12 @@ package needle
 import (
 	"errors"
 	"fmt"
-	"io"
-	"math"
-
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/storage/backend"
 	. "github.com/chrislusf/seaweedfs/weed/storage/types"
 	"github.com/chrislusf/seaweedfs/weed/util"
+	"io"
+	"math"
 )
 
 const (
@@ -156,6 +155,35 @@ func (n *Needle) Append(w backend.BackendStorageFile, version Version) (offset u
 	return offset, size, actualSize, err
 }
 
+func WriteNeedleBlob(w backend.BackendStorageFile, dataSlice []byte, size Size, appendAtNs uint64, version Version) (offset uint64, err error) {
+
+	if end, _, e := w.GetStat(); e == nil {
+		defer func(w backend.BackendStorageFile, off int64) {
+			if err != nil {
+				if te := w.Truncate(end); te != nil {
+					glog.V(0).Infof("Failed to truncate %s back to %d with error: %v", w.Name(), end, te)
+				}
+			}
+		}(w, end)
+		offset = uint64(end)
+	} else {
+		err = fmt.Errorf("Cannot Read Current Volume Position: %v", e)
+		return
+	}
+
+	if version == Version3 {
+		tsOffset := NeedleHeaderSize + size + NeedleChecksumSize
+		util.Uint64toBytes(dataSlice[tsOffset:tsOffset+TimestampSize], appendAtNs)
+	}
+
+	if err == nil {
+		_, err = w.WriteAt(dataSlice, int64(offset))
+	}
+
+	return
+
+}
+
 func ReadNeedleBlob(r backend.BackendStorageFile, offset int64, size Size, version Version) (dataSlice []byte, err error) {
 
 	dataSize := GetActualSize(size, version)

weed/storage/needle_map/memdb.go

@@ -2,6 +2,7 @@ package needle_map
 
 import (
 	"fmt"
+	"io"
 	"os"
 
 	"github.com/syndtr/goleveldb/leveldb"
@@ -104,7 +105,13 @@ func (cm *MemDb) LoadFromIdx(idxName string) (ret error) {
 	}
 	defer idxFile.Close()
 
-	return idx.WalkIndexFile(idxFile, func(key NeedleId, offset Offset, size Size) error {
+	return cm.LoadFromReaderAt(idxFile)
+
+}
+
+func (cm *MemDb) LoadFromReaderAt(readerAt io.ReaderAt) (ret error) {
+
+	return idx.WalkIndexFile(readerAt, func(key NeedleId, offset Offset, size Size) error {
 		if offset.IsZero() || size.IsDeleted() {
 			return cm.Delete(key)
 		}

weed/storage/volume_read.go

@@ -0,0 +1,131 @@
+package storage
+
+import (
+	"fmt"
+	"io"
+	"time"
+
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/storage/backend"
+	"github.com/chrislusf/seaweedfs/weed/storage/needle"
+	"github.com/chrislusf/seaweedfs/weed/storage/super_block"
+	. "github.com/chrislusf/seaweedfs/weed/storage/types"
+)
+
+// read fills in Needle content by looking up n.Id from NeedleMapper
+func (v *Volume) readNeedle(n *needle.Needle, readOption *ReadOption) (int, error) {
+	v.dataFileAccessLock.RLock()
+	defer v.dataFileAccessLock.RUnlock()
+
+	nv, ok := v.nm.Get(n.Id)
+	if !ok || nv.Offset.IsZero() {
+		return -1, ErrorNotFound
+	}
+	readSize := nv.Size
+	if readSize.IsDeleted() {
+		if readOption != nil && readOption.ReadDeleted && readSize != TombstoneFileSize {
+			glog.V(3).Infof("reading deleted %s", n.String())
+			readSize = -readSize
+		} else {
+			return -1, ErrorDeleted
+		}
+	}
+	if readSize == 0 {
+		return 0, nil
+	}
+	err := n.ReadData(v.DataBackend, nv.Offset.ToActualOffset(), readSize, v.Version())
+	if err == needle.ErrorSizeMismatch && OffsetSize == 4 {
+		err = n.ReadData(v.DataBackend, nv.Offset.ToActualOffset()+int64(MaxPossibleVolumeSize), readSize, v.Version())
+	}
+	v.checkReadWriteError(err)
+	if err != nil {
+		return 0, err
+	}
+	bytesRead := len(n.Data)
+	if !n.HasTtl() {
+		return bytesRead, nil
+	}
+	ttlMinutes := n.Ttl.Minutes()
+	if ttlMinutes == 0 {
+		return bytesRead, nil
+	}
+	if !n.HasLastModifiedDate() {
+		return bytesRead, nil
+	}
+	if time.Now().Before(time.Unix(0, int64(n.AppendAtNs)).Add(time.Duration(ttlMinutes) * time.Minute)) {
+		return bytesRead, nil
+	}
+	return -1, ErrorNotFound
+}
+
+// read fills in Needle content by looking up n.Id from NeedleMapper
+func (v *Volume) ReadNeedleBlob(offset int64, size Size) ([]byte, error) {
+	v.dataFileAccessLock.RLock()
+	defer v.dataFileAccessLock.RUnlock()
+
+	return needle.ReadNeedleBlob(v.DataBackend, offset, size, v.Version())
+}
+
+type VolumeFileScanner interface {
+	VisitSuperBlock(super_block.SuperBlock) error
+	ReadNeedleBody() bool
+	VisitNeedle(n *needle.Needle, offset int64, needleHeader, needleBody []byte) error
+}
+
+func ScanVolumeFile(dirname string, collection string, id needle.VolumeId,
+	needleMapKind NeedleMapKind,
+	volumeFileScanner VolumeFileScanner) (err error) {
+	var v *Volume
+	if v, err = loadVolumeWithoutIndex(dirname, collection, id, needleMapKind); err != nil {
+		return fmt.Errorf("failed to load volume %d: %v", id, err)
+	}
+	if err = volumeFileScanner.VisitSuperBlock(v.SuperBlock); err != nil {
+		return fmt.Errorf("failed to process volume %d super block: %v", id, err)
+	}
+	defer v.Close()
+
+	version := v.Version()
+
+	offset := int64(v.SuperBlock.BlockSize())
+
+	return ScanVolumeFileFrom(version, v.DataBackend, offset, volumeFileScanner)
+}
+
+func ScanVolumeFileFrom(version needle.Version, datBackend backend.BackendStorageFile, offset int64, volumeFileScanner VolumeFileScanner) (err error) {
+	n, nh, rest, e := needle.ReadNeedleHeader(datBackend, version, offset)
+	if e != nil {
+		if e == io.EOF {
+			return nil
+		}
+		return fmt.Errorf("cannot read %s at offset %d: %v", datBackend.Name(), offset, e)
+	}
+	for n != nil {
+		var needleBody []byte
+		if volumeFileScanner.ReadNeedleBody() {
+			// println("needle", n.Id.String(), "offset", offset, "size", n.Size, "rest", rest)
+			if needleBody, err = n.ReadNeedleBody(datBackend, version, offset+NeedleHeaderSize, rest); err != nil {
+				glog.V(0).Infof("cannot read needle head [%d, %d) body [%d, %d) body length %d: %v", offset, offset+NeedleHeaderSize, offset+NeedleHeaderSize, offset+NeedleHeaderSize+rest, rest, err)
+				// err = fmt.Errorf("cannot read needle body: %v", err)
+				// return
+			}
+		}
+		err := volumeFileScanner.VisitNeedle(n, offset, nh, needleBody)
+		if err == io.EOF {
+			return nil
+		}
+		if err != nil {
+			glog.V(0).Infof("visit needle error: %v", err)
+			return fmt.Errorf("visit needle error: %v", err)
+		}
+		offset += NeedleHeaderSize + rest
+		glog.V(4).Infof("==> new entry offset %d", offset)
+		if n, nh, rest, err = needle.ReadNeedleHeader(datBackend, version, offset); err != nil {
+			if err == io.EOF {
+				return nil
+			}
+			return fmt.Errorf("cannot read needle header at offset %d: %v", offset, err)
+		}
+		glog.V(4).Infof("new entry needle size:%d rest:%d", n.Size, rest)
+	}
+	return nil
+}

weed/storage/volume_write.go

@@ -4,14 +4,12 @@ import (
 	"bytes"
 	"errors"
 	"fmt"
-	"io"
 	"os"
 	"time"
 
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/storage/backend"
 	"github.com/chrislusf/seaweedfs/weed/storage/needle"
-	"github.com/chrislusf/seaweedfs/weed/storage/super_block"
 	. "github.com/chrislusf/seaweedfs/weed/storage/types"
 )
 
@@ -146,7 +144,8 @@ func (v *Volume) doWriteRequest(n *needle.Needle) (offset uint64, size Size, isU
 			return
 		}
 		if existingNeedle.Cookie != n.Cookie {
-			glog.V(0).Infof("write cookie mismatch: existing %x, new %x", existingNeedle.Cookie, n.Cookie)
+			glog.V(0).Infof("write cookie mismatch: existing %s, new %s",
+				needle.NewFileIdFromNeedle(v.Id, existingNeedle), needle.NewFileIdFromNeedle(v.Id, n))
 			err = fmt.Errorf("mismatching cookie %x", n.Cookie)
 			return
 		}
@@ -226,52 +225,6 @@ func (v *Volume) doDeleteRequest(n *needle.Needle) (Size, error) {
 	return 0, nil
 }
 
-// read fills in Needle content by looking up n.Id from NeedleMapper
-func (v *Volume) readNeedle(n *needle.Needle, readOption *ReadOption) (int, error) {
-	v.dataFileAccessLock.RLock()
-	defer v.dataFileAccessLock.RUnlock()
-
-	nv, ok := v.nm.Get(n.Id)
-	if !ok || nv.Offset.IsZero() {
-		return -1, ErrorNotFound
-	}
-	readSize := nv.Size
-	if readSize.IsDeleted() {
-		if readOption != nil && readOption.ReadDeleted && readSize != TombstoneFileSize {
-			glog.V(3).Infof("reading deleted %s", n.String())
-			readSize = -readSize
-		} else {
-			return -1, ErrorDeleted
-		}
-	}
-	if readSize == 0 {
-		return 0, nil
-	}
-	err := n.ReadData(v.DataBackend, nv.Offset.ToActualOffset(), readSize, v.Version())
-	if err == needle.ErrorSizeMismatch && OffsetSize == 4 {
-		err = n.ReadData(v.DataBackend, nv.Offset.ToActualOffset()+int64(MaxPossibleVolumeSize), readSize, v.Version())
-	}
-	v.checkReadWriteError(err)
-	if err != nil {
-		return 0, err
-	}
-	bytesRead := len(n.Data)
-	if !n.HasTtl() {
-		return bytesRead, nil
-	}
-	ttlMinutes := n.Ttl.Minutes()
-	if ttlMinutes == 0 {
-		return bytesRead, nil
-	}
-	if !n.HasLastModifiedDate() {
-		return bytesRead, nil
-	}
-	if time.Now().Before(time.Unix(0, int64(n.AppendAtNs)).Add(time.Duration(ttlMinutes) * time.Minute)) {
-		return bytesRead, nil
-	}
-	return -1, ErrorNotFound
-}
-
 func (v *Volume) startWorker() {
 	go func() {
 		chanClosed := false
@@ -347,66 +300,28 @@ func (v *Volume) startWorker() {
 	}()
 }
 
-type VolumeFileScanner interface {
-	VisitSuperBlock(super_block.SuperBlock) error
-	ReadNeedleBody() bool
-	VisitNeedle(n *needle.Needle, offset int64, needleHeader, needleBody []byte) error
-}
+func (v *Volume) WriteNeedleBlob(needleId NeedleId, needleBlob []byte, size Size) error {
 
-func ScanVolumeFile(dirname string, collection string, id needle.VolumeId,
-	needleMapKind NeedleMapKind,
-	volumeFileScanner VolumeFileScanner) (err error) {
-	var v *Volume
-	if v, err = loadVolumeWithoutIndex(dirname, collection, id, needleMapKind); err != nil {
-		return fmt.Errorf("failed to load volume %d: %v", id, err)
-	}
-	if err = volumeFileScanner.VisitSuperBlock(v.SuperBlock); err != nil {
-		return fmt.Errorf("failed to process volume %d super block: %v", id, err)
-	}
-	defer v.Close()
-
-	version := v.Version()
+	v.dataFileAccessLock.Lock()
+	defer v.dataFileAccessLock.Unlock()
 
-	offset := int64(v.SuperBlock.BlockSize())
+	if MaxPossibleVolumeSize < v.nm.ContentSize()+uint64(len(needleBlob)) {
+		return fmt.Errorf("volume size limit %d exceeded! current size is %d", MaxPossibleVolumeSize, v.nm.ContentSize())
+	}
 
-	return ScanVolumeFileFrom(version, v.DataBackend, offset, volumeFileScanner)
-}
+	appendAtNs := uint64(time.Now().UnixNano())
+	offset, err := needle.WriteNeedleBlob(v.DataBackend, needleBlob, size, appendAtNs, v.Version())
 
-func ScanVolumeFileFrom(version needle.Version, datBackend backend.BackendStorageFile, offset int64, volumeFileScanner VolumeFileScanner) (err error) {
-	n, nh, rest, e := needle.ReadNeedleHeader(datBackend, version, offset)
-	if e != nil {
-		if e == io.EOF {
-			return nil
-		}
-		return fmt.Errorf("cannot read %s at offset %d: %v", datBackend.Name(), offset, e)
+	v.checkReadWriteError(err)
+	if err != nil {
+		return err
 	}
-	for n != nil {
-		var needleBody []byte
-		if volumeFileScanner.ReadNeedleBody() {
-			// println("needle", n.Id.String(), "offset", offset, "size", n.Size, "rest", rest)
-			if needleBody, err = n.ReadNeedleBody(datBackend, version, offset+NeedleHeaderSize, rest); err != nil {
-				glog.V(0).Infof("cannot read needle head [%d, %d) body [%d, %d) body length %d: %v", offset, offset+NeedleHeaderSize, offset+NeedleHeaderSize, offset+NeedleHeaderSize+rest, rest, err)
-				// err = fmt.Errorf("cannot read needle body: %v", err)
-				// return
-			}
-		}
-		err := volumeFileScanner.VisitNeedle(n, offset, nh, needleBody)
-		if err == io.EOF {
-			return nil
-		}
-		if err != nil {
-			glog.V(0).Infof("visit needle error: %v", err)
-			return fmt.Errorf("visit needle error: %v", err)
-		}
-		offset += NeedleHeaderSize + rest
-		glog.V(4).Infof("==> new entry offset %d", offset)
-		if n, nh, rest, err = needle.ReadNeedleHeader(datBackend, version, offset); err != nil {
-			if err == io.EOF {
-				return nil
-			}
-			return fmt.Errorf("cannot read needle header at offset %d: %v", offset, err)
-		}
-		glog.V(4).Infof("new entry needle size:%d rest:%d", n.Size, rest)
+	v.lastAppendAtNs = appendAtNs
+
+	// add to needle map
+	if err = v.nm.Put(needleId, ToOffset(int64(offset)), size); err != nil {
+		glog.V(4).Infof("failed to put in needle map %d: %v", needleId, err)
 	}
-	return nil
+
+	return err
 }

weed/util/chunk_cache/chunk_cache_in_memory.go

@@ -32,5 +32,7 @@ func (c *ChunkCacheInMemory) GetChunk(fileId string) []byte {
 }
 
 func (c *ChunkCacheInMemory) SetChunk(fileId string, data []byte) {
-	c.cache.Set(fileId, data, time.Hour)
+	localCopy := make([]byte, len(data))
+	copy(localCopy, data)
+	c.cache.Set(fileId, localCopy, time.Hour)
 }

weed/util/constants.go

@@ -5,7 +5,7 @@ import (
 )
 
 var (
-	VERSION = fmt.Sprintf("%s %d.%02d", sizeLimit, 2, 32)
+	VERSION = fmt.Sprintf("%s %d.%02d", sizeLimit, 2, 38)
 	COMMIT  = ""
 )
 

weed/util/fasthttp_util.go

@@ -1,116 +0,0 @@
-package util
-
-import (
-	"bytes"
-	"fmt"
-	"github.com/valyala/fasthttp"
-	"sync"
-	"time"
-)
-
-var (
-	fastClient = &fasthttp.Client{
-		NoDefaultUserAgentHeader:      true, // Don't send: User-Agent: fasthttp
-		MaxConnsPerHost:               1024,
-		ReadBufferSize:                4096,      // Make sure to set this big enough that your whole request can be read at once.
-		WriteBufferSize:               64 * 1024, // Same but for your response.
-		ReadTimeout:                   time.Second,
-		WriteTimeout:                  time.Second,
-		MaxIdleConnDuration:           time.Minute,
-		DisableHeaderNamesNormalizing: true, // If you set the case on your headers correctly you can enable this.
-		DialDualStack:                 true,
-	}
-
-	// Put everything in pools to prevent garbage.
-	bytesPool = sync.Pool{
-		New: func() interface{} {
-			b := make([]byte, 0)
-			return &b
-		},
-	}
-
-	responsePool = sync.Pool{
-		New: func() interface{} {
-			return make(chan *fasthttp.Response)
-		},
-	}
-)
-
-func FastGet(url string) ([]byte, bool, error) {
-
-	req := fasthttp.AcquireRequest()
-	res := fasthttp.AcquireResponse()
-	defer fasthttp.ReleaseRequest(req)
-	defer fasthttp.ReleaseResponse(res)
-
-	req.SetRequestURIBytes([]byte(url))
-	req.Header.Add("Accept-Encoding", "gzip")
-
-	err := fastClient.Do(req, res)
-	if err != nil {
-		return nil, true, err
-	}
-
-	var data []byte
-	contentEncoding := res.Header.Peek("Content-Encoding")
-	if bytes.Compare(contentEncoding, []byte("gzip")) == 0 {
-		data, err = res.BodyGunzip()
-	} else {
-		data = res.Body()
-	}
-
-	out := make([]byte, len(data))
-	copy(out, data)
-
-	if res.StatusCode() >= 400 {
-		retryable := res.StatusCode() >= 500
-		return nil, retryable, fmt.Errorf("%s: %d", url, res.StatusCode())
-	}
-	if err != nil {
-		return nil, false, err
-	}
-	return out, false, nil
-}
-
-func FastReadUrlAsStream(fileUrl string, cipherKey []byte, isContentGzipped bool, isFullChunk bool, offset int64, size int, fn func(data []byte)) (retryable bool, err error) {
-
-	if cipherKey != nil {
-		return readEncryptedUrl(fileUrl, cipherKey, isContentGzipped, isFullChunk, offset, size, fn)
-	}
-
-	req := fasthttp.AcquireRequest()
-	res := fasthttp.AcquireResponse()
-	defer fasthttp.ReleaseRequest(req)
-	defer fasthttp.ReleaseResponse(res)
-
-	req.SetRequestURIBytes([]byte(fileUrl))
-
-	if isFullChunk {
-		req.Header.Add("Accept-Encoding", "gzip")
-	} else {
-		req.Header.Add("Range", fmt.Sprintf("bytes=%d-%d", offset, offset+int64(size)-1))
-	}
-
-	if err = fastClient.Do(req, res); err != nil {
-		return true, err
-	}
-
-	if res.StatusCode() >= 400 {
-		retryable = res.StatusCode() >= 500
-		return retryable, fmt.Errorf("%s: %d", fileUrl, res.StatusCode())
-	}
-
-	contentEncoding := res.Header.Peek("Content-Encoding")
-	if bytes.Compare(contentEncoding, []byte("gzip")) == 0 {
-		bodyData, err := res.BodyGunzip()
-		if err != nil {
-			return false, err
-		}
-		fn(bodyData)
-	} else {
-		fn(res.Body())
-	}
-
-	return false, nil
-
-}

weed/util/http_util.go

@@ -313,7 +313,7 @@ func ReadUrlAsStream(fileUrl string, cipherKey []byte, isContentGzipped bool, is
 }
 
 func readEncryptedUrl(fileUrl string, cipherKey []byte, isContentCompressed bool, isFullChunk bool, offset int64, size int, fn func(data []byte)) (bool, error) {
-	encryptedData, retryable, err := FastGet(fileUrl)
+	encryptedData, retryable, err := Get(fileUrl)
 	if err != nil {
 		return retryable, fmt.Errorf("fetch %s: %v", fileUrl, err)
 	}

weed/util/limiter.go

@@ -1,5 +1,12 @@
 package util
 
+import (
+	"math/rand"
+	"reflect"
+	"sync"
+	"sync/atomic"
+)
+
 // initial version comes from https://github.com/korovkin/limiter/blob/master/limiter.go
 
 // LimitedConcurrentExecutor object
@@ -38,3 +45,70 @@ func (c *LimitedConcurrentExecutor) Execute(job func()) {
 		job()
 	}()
 }
+
+// a different implementation, but somehow more "conservative"
+type OperationRequest func()
+
+type LimitedOutOfOrderProcessor struct {
+	processorSlots     uint32
+	processors         []chan OperationRequest
+	processorLimit     int32
+	processorLimitCond *sync.Cond
+	currentProcessor   int32
+}
+
+func NewLimitedOutOfOrderProcessor(limit int32) (c *LimitedOutOfOrderProcessor) {
+
+	processorSlots := uint32(32)
+	c = &LimitedOutOfOrderProcessor{
+		processorSlots:     processorSlots,
+		processors:         make([]chan OperationRequest, processorSlots),
+		processorLimit:     limit,
+		processorLimitCond: sync.NewCond(new(sync.Mutex)),
+	}
+
+	for i := 0; i < int(processorSlots); i++ {
+		c.processors[i] = make(chan OperationRequest)
+	}
+
+	cases := make([]reflect.SelectCase, processorSlots)
+	for i, ch := range c.processors {
+		cases[i] = reflect.SelectCase{Dir: reflect.SelectRecv, Chan: reflect.ValueOf(ch)}
+	}
+
+	go func() {
+		for {
+			_, value, ok := reflect.Select(cases)
+			if !ok {
+				continue
+			}
+
+			request := value.Interface().(OperationRequest)
+
+			if c.processorLimit > 0 {
+				c.processorLimitCond.L.Lock()
+				for atomic.LoadInt32(&c.currentProcessor) > c.processorLimit {
+					c.processorLimitCond.Wait()
+				}
+				atomic.AddInt32(&c.currentProcessor, 1)
+				c.processorLimitCond.L.Unlock()
+			}
+
+			go func() {
+				if c.processorLimit > 0 {
+					defer atomic.AddInt32(&c.currentProcessor, -1)
+					defer c.processorLimitCond.Signal()
+				}
+				request()
+			}()
+
+		}
+	}()
+
+	return c
+}
+
+func (c *LimitedOutOfOrderProcessor) Execute(request OperationRequest) {
+	index := rand.Uint32() % c.processorSlots
+	c.processors[index] <- request
+}