Chris LuandGitHub
e377149d39
* mount: add the WinFsp filesystem adapter WinFsp speaks a path-based FUSE dialect; weed/mount implements the inode-based raw protocol the Linux kernel uses. This translates between them so Windows runs the same filesystem code as everywhere else rather than a second implementation: paths resolve to inodes one Lookup at a time, and the raw operations run unchanged underneath. Errno translation is spelled out rather than passed through. Go numbers Windows errnos as offsets from APPLICATION_ERROR, so the raw value would mean something unrelated by the time WinFsp read it. Hard links return ENOSYS since WinFsp has none, and byte-range locks stay with its kernel driver rather than the mount's lock table. Not reachable from the mount command yet. * mount: build the winfsp errno table with explicit precedence Platforms alias errnos differently: freebsd has no ENODATA and linux makes ENOATTR the same value as it. A map literal with colliding constant keys does not compile, so build the table and let the first entry win, keeping the general codes their own meaning. * mount: wire the winfsp adapter into the mount command RunMount was one function doing filer setup, mount-point preparation and serving. The setup is the same everywhere, so it moves to mount_common.go and each platform keeps only what differs. Windows differs mostly in the mount point: WinFsp wants a drive letter or a path that does not exist yet, so none of the unix preparation applies, and a bad one is worth rejecting up front because WinFsp reports failure as a bare false. Adds -windows.caseInsensitive for software that expects Windows naming rules. * ci: mount on windows and exercise it Builds weed.exe, installs WinFsp, starts a cluster, mounts S: and runs a test suite against it: round trips at several sizes, offset writes, rename, delete, nested directories, concurrent writers, and a directory wide enough to stand in for the case that prompted this. Nothing else here can run the Windows mount, so without this the adapter is only known to compile. * ci: build the windows mount without cgo The runner has MinGW, so cgo is on by default and cgofuse compiles its cgo variant, which needs WinFsp's headers. The nocgo variant loads the DLL at run time and is what the released weed.exe uses. * mount: make the winfsp path splitting portable and test it resolve and resolveParent had the splitting inline in a windows-tagged file, so the cases that matter most there — both separators, empty and dot components, the root having no parent to create in — could not be tested on any runner that builds this. * test: check the windows mount persists across a remount Reading a file back through the same live mount proves nothing about durability; the answer can come from the mount's own caches. Write the fixtures, confirm the filer serves them with the mount out of the path, then re-read after a teardown and remount. * test: cover the windows mount operations that had none Truncate, append, chtimes and the hard-link refusal were implemented but never exercised, and the errno table was only unit-tested for mapping, never end to end. Adds names that have to survive the UTF-16 boundary, rename over an existing target and across directories, and concurrent handles on one file rather than one file each. * ci: dial the filer over ipv4 and run the persistence phases localhost resolves to ::1 first on windows and the cluster binds ipv4 only, so the mount's grpc dial was refused while the http readiness probe passed by falling back to ipv4. * ci: pin the cluster to loopback and probe ports by connecting weed mini advertises the runner's LAN address and binds filer grpc there, so the mount's dial to 127.0.0.1:18888 was refused while http answered. The readiness probe also passed with nothing on 18888: Test-NetConnection reported success for a port that then refused a connection, so it now opens a socket instead. * ci: report listening ports before mounting The readiness probe connects to the filer grpc port and the mount is then refused on it, which cannot both be true; print the actual state. * ci: run the cluster, mount and tests in one step The runner tears down a step's process tree when its shell exits, so the cluster started in an earlier step was already gone: the readiness probe passed against a live filer, the step ended, and the mount then found nothing listening. A diagnostic step reported no weed.exe at all. Everything that needs those processes alive now shares a step. * mount: key windows file io on the handle, not the path Read and Write walked the path on every call to fill in a NodeId the raw filesystem never reads: both look the file up by handle. Under eight writers creating files in one directory the walk transiently missed and the write failed with ENOENT before reaching the filesystem at all. Same for flush, fsync and the release calls. O_EXCL now fails on an existing name instead of taking it over, and Symlink is refused: the entry is easy to create but WinFsp only follows it once the reparse point is wired up, so it read back as an empty file. * mount: translate cgofuse open flags for windows cgofuse reports MSVC's numbering and the raw filesystem tests Go's, so only the access mode and O_TRUNC lined up: O_EXCL arrived as O_APPEND and O_CREAT as nothing at all. Also report which handle a failed write was using, to tell a handle that was never issued from one released while still in use. * mount: report which step of a windows create failed A concurrent create fails with ENOENT and the path walk, the parent lookup and the create itself are indistinguishable from the caller. * ci: send weed logs to stderr on windows glog writes to its own files by default, so the mount's own error output never reached the redirected log. Its flags are global and have to come before the subcommand. * mount: resolve known paths from the inode table on windows Every create walked the parent chain with a filer lookup per component. With eight writers creating files in one directory that is hundreds of concurrent lookups of the same parent, and lookupEntry reports an authoritative ENOENT when the directory is cached, the entry is not in the cache and the inode table has no record — a window a concurrent refresh can open for a directory that plainly exists. A path the mount already tracks now resolves straight out of that table. * test: sync the windows persistence fixtures before closing The mount is killed rather than unmounted, so anything still queued for flush is legitimately lost and the test was measuring crash durability while calling it persistence. A 9MB file lost four chunks that way. * mount: keep the lookup refresh on the target path Resolving a tracked path straight from the inode table skipped Lookup, which is also what refreshes the entry: a truncate then read back the pre-truncate size. Only the parent chain takes the shortcut now, which is where the concurrent creates were racing anyway. * mount: log every windows resolve failure Open suppressed ENOENT and Getattr logged nothing, which hid the two callbacks that can report a missing file during a create. * mount: drop dot entries from windows directory listings readdir reports "." and ".." for the kernel, but Windows enumerates a directory without them and displays whatever it is handed, so a folder of 200 files listed 202. Go's ReadDir filters them, which is why only the PowerShell walk caught it. * mount: flush queued writes when windows mount is interrupted The signal handler exits the process the moment its hooks return, so the WaitForAsyncFlush after Serve never ran on ctrl-c and queued writes were dropped. * mount: let windows mount over an empty directory WinFsp turns a directory mount point into a reparse point, which NTFS allows on an empty directory and refuses on a populated one. The check rejected every existing directory, so the ordinary habit of creating the mount point first failed with a message saying it should not exist. CI now mounts over a pre-created directory and writes through it. * ci: run the windows mount check on any pull request It is the only thing that exercises the Windows mount, so restricting it to pull requests based on master skipped it for stacked ones. Replaces the branch name that was pushed to trigger it. * mount: do not log a missing windows entry as an error Windows probes for entries that do not exist as a matter of course, so ENOENT from getattr and open is an answer rather than a fault and would have filled the log. * mount: take the fast path for parent chains in every windows resolve Narrowing it to resolveParent left Getattr and Open re-walking the parent with a filer lookup per component, and those are what Windows calls before a create: eight writers in one directory still raced a meta cache refresh there. Only the final component needs the Lookup refresh. The pass that suggested otherwise came from a run five times slower than the failing ones, where the race had no room to appear. * mount: drop the windows path resolution shortcut Resolving from the inode table skipped the Lookup that refreshes an entry, and a truncate then read back its old size. Applying it only to the parent chain kept truncate correct but left concurrent creates failing, and applying it to the final component too inverted that. The two cannot both be satisfied this way, so this returns to looking up every component and leaves the concurrent create failure open. * mount: fall back to the open handle when a deferred entry is evicted A create that defers the filer write leaves the entry only in the local cache. Creating many files at once pushes the directory past the hot threshold and evicts it, taking that placeholder with it, so a lookup went to the filer, found nothing, and reported a file that plainly exists as missing. The handle still holding the unflushed entry is authoritative for it. Caught by concurrent creates over a Windows mount, which resolves a path on every call rather than relying on a kernel dentry cache. * mount: let cgofuse resolve to the version the module graph requires rclone already depends on cgofuse at a newer commit than the v1.6.0 pin, so readonly builds refused the go.mod until it matched what MVS picks. The interface and flag values the adapter uses are unchanged there. * mount: wait for a pending async flush before looking up on the filer Open, unlink and rename already wait, but a plain lookup went straight to the filer and read pre-close metadata: truncate a file, close it, and a path probe during the flush window reported the old size. The kernel attr cache hides this on linux; a front end that resolves paths on every operation hit it directly. * mount: reject a umask wider than the file mode it becomes ParseUint allowed 64 bits and the result is narrowed to os.FileMode, which is 32, so an out-of-range umask truncated silently instead of being reported as unparseable. * mount: address review findings on the windows mount WaitForAsyncFlush closed its channel unconditionally and shutdown reaches it from both the interrupt hook and the path that resumes after serving, so a ctrl-c could panic on a second close. The deferred-entry fallback read an open handle's entry without its lock, which is what the other two readers of that field take so FromPbEntry does not walk the chunk slice mid-append. The async-flush wait also sat ahead of the meta cache, making every stat of a recently closed file queue behind uploads; it belongs just before the filer is consulted. Windows entries were persisted as uid 0: the raw filesystem stores InHeader's owner and the adapter left it zero. They now carry the identity the mount was started with. The errno table used Linux numbering while cgofuse decodes MSVC's, so ENAMETOOLONG arrived as EDEADLK and five others were likewise wrong; a windows test pins each value to cgofuse's own constant. Also: break the filer handshake loop on success rather than always running ten rounds, accept a drive letter written S:\\, report a missing WinFsp instead of panicking, keep commas out of the volume label, and drop -windows.caseInsensitive, which told WinFsp the mount folds case while lookups stayed exact. * mount: return windows lookup references so the inode table stays bounded Every operation that hands back an EntryOut grants a reference the Linux kernel returns with FORGET. WinFsp has no FORGET, so the adapter took one per path component per call, plus one per child of every readdirplus, and never gave any back: inodeToPath grew for the life of the mount. Walking the 200k-file directory this exists for stranded 200k references. The adapter now plays the part the kernel plays. Each resolution releases what it took, and an open handle keeps the reference for its inode until Release, counted because the raw filesystem reuses one handle for repeated opens. Holding it is not optional: completeAsyncFlush skips the metadata flush when the saved path no longer maps to the inode, so releasing early would lose a close's metadata. Also stops persisting the display owner. -o uid=-1 makes WinFsp report the calling user whatever we say, but the value handed to the raw filesystem is written to the filer, and 4294967295 is what every other client would read. -windows.uid and -windows.gid set what is recorded. * mount: fix windows behaviours the reference implementations guard against WinFsp has no ro option — it discards the flag and leaves the volume writable — so -readOnly accepted writes and deletes. The refusal now happens in the operations themselves. Windows sends times around its own 1601 epoch, which arrive as a large negative second count; casting them through stored a year-1601 timestamp that every other client then read. Those are now left alone. rclone carries the same guard. Chown returned ENOSYS, and WinFsp passes a chown failure straight out of SetSecurity, so Explorer's Security tab and icacls failed for edits that were not about ownership. It now accepts and discards. Only create and mkdir presented a caller; the rest sent uid 0, which hasAccess treats as root, so deletes and renames skipped the permission check that creates got. Every operation presents the same identity now. A drive letter written S:\ reached WinFsp unnormalised, which recognises a drive only as exactly two characters and then failed as a directory path. A test also pins the open flag translation, since swapping O_EXCL and O_TRUNC would turn 'fail if it exists' into 'truncate it'. * mount: answer windows getattr and truncate from the open handle WinFsp keeps the path a handle was opened with and never updates it when the file is renamed, so resolving the path again fails on a handle that is still perfectly valid — the ordinary write-temp-then-rename save pattern. The handle already knows its inode, which also removes a full path walk from two operations WinFsp calls constantly. Readlink on the root now refuses. WinFsp probes there to decide whether the volume has symlinks and enables them unless it fails, and with them on it resolves a path a component at a time, each one reaching us as its own walk — all for a feature Symlink already refuses. * mount: require the windows mount directory not to exist WinFsp creates the directory itself with FILE_CREATE and removes it when the filesystem goes away, so an existing one — empty or not — fails with "mount point in use". Allowing an empty directory was wrong, and the CI check that appeared to prove otherwise was the vacuous one: listing a plain directory succeeds whether or not anything is mounted on it, so the step passed while the mount had failed and the writes went to local disk. That check now waits for the reparse point, which is what caught this. * mount: apply review comments on the windows mount -windows.uid and -windows.gid reached the adapter but not the filesystem parameters, which is what carries the owner written to the filer, so the flags changed nothing. Readdir re-resolved the path while Getattr and Truncate answer from the handle; a directory renamed during an enumeration then failed on the stale path WinFsp still holds. Utimens now honours UTIME_OMIT instead of writing whatever came with it. * mount: tag the unix-only lock tests away from windows The production lock files were tagged when the package was made to build on windows, but the tests that exercise them were not, so anything that compiles tests for windows still failed on syscall.F_WRLCK. * ci: vet the mount tests for each target too Only compiling the non-test build let an untagged test keep a per-OS syscall constant without anything noticing.
SeaweedFS
Sponsor SeaweedFS via Patreon
SeaweedFS is an independent Apache-licensed open source project with its ongoing development made possible entirely thanks to the support of these awesome backers. If you'd like to grow SeaweedFS even stronger, please consider joining our sponsors on Patreon.
Your support will be really appreciated by me and other supporters!
Gold Sponsors
- Download Binaries for different platforms
- SeaweedFS on Slack
- SeaweedFS on Twitter
- SeaweedFS on Telegram
- SeaweedFS on Reddit
- SeaweedFS Mailing List
- Wiki Documentation
- SeaweedFS White Paper
- SeaweedFS Introduction Slides 2025.5
- SeaweedFS Introduction Slides 2021.5
- SeaweedFS Introduction Slides 2019.3
Table of Contents
- Quick Start
- Introduction
- Features
- Example: Using Seaweed Object Store
- Architecture
- Compared to Other File Systems
- Dev Plan
- Installation Guide
- Disk Related Topics
- Benchmark
- Enterprise
- License
Quick Start
Quick Start with weed mini
Download the latest binary from https://github.com/seaweedfs/seaweedfs/releases and unzip the single weed (or weed.exe) file, or run go install github.com/seaweedfs/seaweedfs/weed@latest. Then start a ready-to-use S3 object store with credentials and a pre-created bucket in one command:
AWS_ACCESS_KEY_ID=admin \
AWS_SECRET_ACCESS_KEY=secret \
S3_BUCKET=my-bucket \
./weed mini -dir=/data
That's it — the S3 endpoint is at http://localhost:8333, my-bucket already exists, and admin/secret are valid credentials. S3_BUCKET accepts a comma-separated list (e.g. raw,processed); use S3_TABLE_BUCKET for S3 Tables (Iceberg) buckets. Drop any of the env vars to skip that piece (no AWS keys → S3 runs in unauthenticated "Allow All" mode for development).
The same command starts everything else too:
- S3 Endpoint: http://localhost:8333
- Master UI: http://localhost:9333
- Volume Server: http://localhost:9340
- Filer UI: http://localhost:8888
- WebDAV: http://localhost:7333
- Admin UI: http://localhost:23646
macOS: if the binary is quarantined, run
xattr -d com.apple.quarantine ./weedfirst.
Perfect for development, testing, learning SeaweedFS, and single-node deployments. To scale out, add more volume servers by running weed volume -dir="/some/data/dir2" -master="<master_host>:9333" -port=8081 locally, on another machine, or on thousands of machines.
Quick Start for S3 API on Docker
docker run -p 8333:8333 \
-e AWS_ACCESS_KEY_ID=admin \
-e AWS_SECRET_ACCESS_KEY=secret \
-e S3_BUCKET=my-bucket \
chrislusf/seaweedfs
Same behavior as the weed mini command above — the S3 endpoint is at http://localhost:8333 with my-bucket pre-created. Drop the env vars to run anonymously for development.
Introduction
SeaweedFS is a simple and highly scalable distributed file system. There are two objectives:
- to store billions of files!
- to serve the files fast!
SeaweedFS started as a blob store to handle small files efficiently. Instead of managing all file metadata in a central master, the central master only manages volumes on volume servers, and these volume servers manage files and their metadata. This relieves concurrency pressure from the central master and spreads file metadata into volume servers, allowing faster file access (O(1), usually just one disk read operation).
There is only 40 bytes of disk storage overhead for each file's metadata. It is so simple with O(1) disk reads that you are welcome to challenge the performance with your actual use cases.
SeaweedFS started by implementing Facebook's Haystack design paper. Also, SeaweedFS implements erasure coding with ideas from f4: Facebook’s Warm BLOB Storage System, and has a lot of similarities with Facebook’s Tectonic Filesystem and Google's Colossus File System
On top of the blob store, optional Filer can support directories and POSIX attributes. Filer is a separate linearly-scalable stateless server with customizable metadata stores, e.g., MySql, Postgres, Redis, Cassandra, HBase, Mongodb, Elastic Search, LevelDB, RocksDB, Sqlite, MemSql, TiDB, Etcd, CockroachDB, YDB, etc.
SeaweedFS can transparently integrate with the cloud. With hot data on local cluster, and warm data on the cloud with O(1) access time, SeaweedFS can achieve both fast local access time and elastic cloud storage capacity. What's more, the cloud storage access API cost is minimized. Faster and cheaper than direct cloud storage!
SeaweedFS also ships a built-in Iceberg REST Catalog, turning the same cluster into a self-contained lakehouse. Spark, Trino, Dremio, DuckDB, and RisingWave can query Iceberg tables directly — no Hive Metastore, Glue, or external catalog service required. Storage and table metadata live in one system, simplifying on-prem and small-team analytics stacks.
Features
Additional Blob Store Features
- Support different replication levels, with rack and data center aware.
- Automatic master servers failover - no single point of failure (SPOF).
- Automatic compression depending on file MIME type.
- Automatic compaction to reclaim disk space after deletion or update.
- Automatic entry TTL expiration.
- Flexible Capacity Expansion: Any server with some disk space can add to the total storage space.
- Adding/Removing servers does not cause any data re-balancing unless triggered by admin commands.
- Optional picture resizing.
- Support ETag, Accept-Range, Last-Modified, etc.
- Support in-memory/leveldb/readonly mode tuning for memory/performance balance.
- Support rebalancing the writable and readonly volumes.
- Customizable Multiple Storage Tiers: Customizable storage disk types to balance performance and cost.
- Transparent cloud integration: unlimited capacity via tiered cloud storage for warm data.
- Erasure Coding for warm storage Rack-Aware 10.4 erasure coding reduces storage cost and increases availability. Enterprise version can customize EC ratio.
Filer Features
- Filer server provides "normal" directories and files via HTTP.
- File TTL automatically expires file metadata and actual file data.
- Mount filer reads and writes files directly as a local directory via FUSE.
- Filer Store Replication enables HA for filer meta data stores.
- Active-Active Replication enables asynchronous one-way or two-way cross cluster continuous replication.
- Amazon S3 compatible API accesses files with S3 tooling.
- Hadoop Compatible File System accesses files from Hadoop/Spark/Flink/etc or even runs HBase.
- Async Replication To Cloud has extremely fast local access and backups to Amazon S3, Google Cloud Storage, Azure, BackBlaze.
- WebDAV accesses as a mapped drive on Mac and Windows, or from mobile devices.
- AES256-GCM Encrypted Storage safely stores the encrypted data.
- Super Large Files stores large or super large files in tens of TB.
- Cloud Drive mounts cloud storage to local cluster, cached for fast read and write with asynchronous write back.
- Gateway to Remote Object Store mirrors bucket operations to remote object storage, in addition to Cloud Drive
Data Lakehouse Features
- S3 Table Buckets expose a dedicated namespace for Iceberg tables with strict layout validation.
- Built-in Iceberg REST Catalog runs alongside the S3 endpoint — no external metastore needed.
- Native integrations with Apache Spark, Trino, Dremio, DuckDB, and RisingWave.
- Automated table maintenance: compaction, snapshot expiration, orphan removal, manifest rewriting.
- Granular IAM at the bucket, namespace, and table level via standard S3 bucket policies.
Kubernetes
- Kubernetes CSI Driver A Container Storage Interface (CSI) Driver.
- SeaweedFS Operator
Example: Using Seaweed Blob Store
By default, the master node runs on port 9333, and the volume nodes run on port 8080. Let's start one master node, and two volume nodes on port 8080 and 8081. Ideally, they should be started from different machines. We'll use localhost as an example.
SeaweedFS uses HTTP REST operations to read, write, and delete. The responses are in JSON or JSONP format.
Start Master Server
> ./weed master
Start Volume Servers
> weed volume -dir="/tmp/data1" -max=5 -master="localhost:9333" -port=8080 &
> weed volume -dir="/tmp/data2" -max=10 -master="localhost:9333" -port=8081 &
Write A Blob
A blob, also referred as a needle, a chunk, or mistakenly as a file, is just a byte array. It can have attributes, such as name, mime type, create or update time, etc. But basically it is just a byte array of a relatively small size, such as 2 MB ~ 64 MB. The size is not fixed.
To upload a blob: first, send a HTTP POST, PUT, or GET request to /dir/assign to get an fid and a volume server URL:
> curl http://localhost:9333/dir/assign
{"count":1,"fid":"3,01637037d6","url":"127.0.0.1:8080","publicUrl":"localhost:8080"}
Second, to store the blob content, send a HTTP multi-part POST request to url + '/' + fid from the response:
> curl -F file=@/home/chris/myphoto.jpg http://127.0.0.1:8080/3,01637037d6
{"name":"myphoto.jpg","size":43234,"eTag":"1cc0118e"}
To update, send another POST request with updated blob content.
For deletion, send an HTTP DELETE request to the same url + '/' + fid URL:
> curl -X DELETE http://127.0.0.1:8080/3,01637037d6
Save Blob Id
Now, you can save the fid, 3,01637037d6 in this case, to a database field.
The number 3 at the start represents a volume id. After the comma, it's one file key, 01, and a file cookie, 637037d6.
The volume id is an unsigned 32-bit integer. The file key is an unsigned 64-bit integer. The file cookie is an unsigned 32-bit integer, used to prevent URL guessing.
The file key and file cookie are both coded in hex. You can store the <volume id, file key, file cookie> tuple in your own format, or simply store the fid as a string.
If stored as a string, in theory, you would need 8+1+16+8=33 bytes. A char(33) would be enough, if not more than enough, since most uses will not need 2^32 volumes.
If space is really a concern, you can store the file id in the binary format. You would need one 4-byte integer for volume id, 8-byte long number for file key, and a 4-byte integer for the file cookie. So 16 bytes are more than enough.
Read a Blob
Here is an example of how to render the URL.
First look up the volume server's URLs by the file's volumeId:
> curl http://localhost:9333/dir/lookup?volumeId=3
{"volumeId":"3","locations":[{"publicUrl":"localhost:8080","url":"localhost:8080"}]}
Since (usually) there are not too many volume servers, and volumes don't move often, you can cache the results most of the time. Depending on the replication type, one volume can have multiple replica locations. Just randomly pick one location to read.
Now you can take the public URL, render the URL or directly read from the volume server via URL:
http://localhost:8080/3,01637037d6.jpg
Notice we add a file extension ".jpg" here. It's optional and just one way for the client to specify the file content type.
If you want a nicer URL, you can use one of these alternative URL formats:
http://localhost:8080/3/01637037d6/my_preferred_name.jpg
http://localhost:8080/3/01637037d6.jpg
http://localhost:8080/3,01637037d6.jpg
http://localhost:8080/3/01637037d6
http://localhost:8080/3,01637037d6
If you want to get a scaled version of an image, you can add some params:
http://localhost:8080/3/01637037d6.jpg?height=200&width=200
http://localhost:8080/3/01637037d6.jpg?height=200&width=200&mode=fit
http://localhost:8080/3/01637037d6.jpg?height=200&width=200&mode=fill
Rack-Aware and Data Center-Aware Replication
SeaweedFS applies the replication strategy at a volume level. So, when you are getting a blob id, you can specify the replication strategy. For example:
curl http://localhost:9333/dir/assign?replication=001
The replication parameter options are:
000: no replication
001: replicate once on the same rack
010: replicate once on a different rack, but same data center
100: replicate once on a different data center
200: replicate twice on two different data center
110: replicate once on a different rack, and once on a different data center
More details about replication can be found on the wiki.
You can also set the default replication strategy when starting the master server.
Allocate Blob Key on Specific Data Center
Volume servers can be started with a specific data center name:
weed volume -dir=/tmp/1 -port=8080 -dataCenter=dc1
weed volume -dir=/tmp/2 -port=8081 -dataCenter=dc2
When requesting a blob key, an optional "dataCenter" parameter can limit the assigned volume to the specific data center. For example, this specifies that the assigned volume should be limited to 'dc1':
http://localhost:9333/dir/assign?dataCenter=dc1
Other Features
- No Single Point of Failure
- Insert with your own keys
- Chunking large files
- Collection as a Simple Name Space
Blob Store Architecture
Usually distributed file systems split each file into chunks. A central server keeps a mapping of filenames to chunks, and also which chunks each chunk server has.
The main drawback is that the central server can't handle many small files efficiently, and since all read requests need to go through the central master, so it might not scale well for many concurrent users.
Instead of managing chunks, SeaweedFS manages data volumes in the master server. Each data volume is 32GB in size, and can hold a lot of blobs. And each storage node can have many data volumes. So the master node only needs to store the metadata about the volumes, which is a fairly small amount of data and is generally stable.
The actual blob metadata, which are the blob volume, offset, and size, is stored in each volume on volume servers. Since each volume server only manages metadata of blobs on its own disk, with only 16 bytes for each blob, all access can read the metadata just from memory and only needs one disk operation to actually read file data.
For comparison, consider that an xfs inode structure in Linux is 536 bytes.
Master Server and Volume Server
The architecture is fairly simple. The actual data is stored in volumes on storage nodes. One volume server can have multiple volumes, and can both support read and write access with basic authentication.
All volumes are managed by a master server. The master server contains the volume id to volume server mapping. This is fairly static information, and can be easily cached.
On each write request, the master server also generates a file key, which is a growing 64-bit unsigned integer. Since write requests are not generally as frequent as read requests, one master server should be able to handle the concurrency well.
Write and Read files
When a client sends a write request, the master server returns (volume id, file key, file cookie, volume node URL) for the blob. The client then contacts the volume node and POSTs the blob content.
When a client needs to read a blob based on (volume id, file key, file cookie), it asks the master server by the volume id for the (volume node URL, volume node public URL), or retrieves this from a cache. Then the client can GET the content, or just render the URL on web pages and let browsers fetch the content.
Saving memory
All blob metadata stored on a volume server is readable from memory without disk access. Each file takes just a 16-byte map entry of <64bit key, 32bit offset, 32bit size>. Of course, each map entry has its own space cost for the map. But usually the disk space runs out before the memory does.
Tiered Storage to the cloud
The local volume servers are much faster, while cloud storages have elastic capacity and are actually more cost-efficient if not accessed often (usually free to upload, but relatively costly to access). With the append-only structure and O(1) access time, SeaweedFS can take advantage of both local and cloud storage by offloading the warm data to the cloud.
Usually hot data are fresh and warm data are old. SeaweedFS puts the newly created volumes on local servers, and optionally upload the older volumes on the cloud. If the older data are accessed less often, this literally gives you unlimited capacity with limited local servers, and still fast for new data.
With the O(1) access time, the network latency cost is kept at minimum.
If the hot/warm data is split as 20/80, with 20 servers, you can achieve storage capacity of 100 servers. That's a cost saving of 80%! Or you can repurpose the 80 servers to store new data also, and get 5X storage throughput.
SeaweedFS Filer
Built on top of the blob store, SeaweedFS Filer adds directory structure to create a file system. The directory structure is an interface that is implemented in many key-value stores or databases.
The content of a file is mapped to one or many blobs, distributed to multiple volumes on multiple volume servers.
Compared to Other File Systems
Most other distributed file systems seem more complicated than necessary.
SeaweedFS is meant to be fast and simple, in both setup and operation. If you do not understand how it works when you reach here, we've failed! Please raise an issue with any questions or update this file with clarifications.
SeaweedFS is constantly moving forward. Same with other systems. These comparisons can be outdated quickly. Please help to keep them updated.
Compared to HDFS
HDFS uses the chunk approach for each file, and is ideal for storing large files.
SeaweedFS is ideal for serving relatively smaller files quickly and concurrently.
SeaweedFS can also store extra large files by splitting them into manageable data chunks, and store the file ids of the data chunks into a meta chunk. This is managed by "weed upload/download" tool, and the weed master or volume servers are agnostic about it.
Compared to GlusterFS, Ceph
The architectures are mostly the same. SeaweedFS aims to store and read files fast, with a simple and flat architecture. The main differences are
- SeaweedFS optimizes for small files, ensuring O(1) disk seek operation, and can also handle large files.
- SeaweedFS statically assigns a volume id for a file. Locating file content becomes just a lookup of the volume id, which can be easily cached.
- SeaweedFS Filer metadata store can be any well-known and proven data store, e.g., Redis, Cassandra, HBase, Mongodb, Elastic Search, MySql, Postgres, Sqlite, MemSql, TiDB, CockroachDB, Etcd, YDB etc, and is easy to customize.
- SeaweedFS Volume server also communicates directly with clients via HTTP, supporting range queries, direct uploads, etc.
| System | File Metadata | File Content Read | POSIX | REST API | Optimized for large number of small files |
|---|---|---|---|---|---|
| SeaweedFS | lookup volume id, cacheable | O(1) disk seek | Yes | Yes | |
| SeaweedFS Filer | Linearly Scalable, Customizable | O(1) disk seek | FUSE | Yes | Yes |
| GlusterFS | hashing | FUSE, NFS | |||
| Ceph | hashing + rules | FUSE | Yes | ||
| MooseFS | in memory | FUSE | No | ||
| MinIO | separate meta file for each file | Yes | No |
Compared to GlusterFS
GlusterFS stores files, both directories and content, in configurable volumes called "bricks".
GlusterFS hashes the path and filename into ids, and assigned to virtual volumes, and then mapped to "bricks".
Compared to MooseFS
MooseFS chooses to neglect small file issue. From moosefs 3.0 manual, "even a small file will occupy 64KiB plus additionally 4KiB of checksums and 1KiB for the header", because it "was initially designed for keeping large amounts (like several thousands) of very big files"
MooseFS Master Server keeps all meta data in memory. Same issue as HDFS namenode.
Compared to Ceph
Ceph can be setup similar to SeaweedFS as a key->blob store. It is much more complicated, with the need to support layers on top of it. Here is a more detailed comparison
SeaweedFS has a centralized master group to look up free volumes, while Ceph uses hashing and metadata servers to locate its objects. Having a centralized master makes it easy to code and manage.
Ceph, like SeaweedFS, is based on the object store RADOS. Ceph is rather complicated with mixed reviews.
Ceph uses CRUSH hashing to automatically manage data placement, which is efficient to locate the data. But the data has to be placed according to the CRUSH algorithm. Any wrong configuration would cause data loss. Topology changes, such as adding new servers to increase capacity, will cause data migration with high IO cost to fit the CRUSH algorithm. SeaweedFS places data by assigning them to any writable volumes. If writes to one volume failed, just pick another volume to write. Adding more volumes is also as simple as it can be.
SeaweedFS is optimized for small files. Small files are stored as one continuous block of content, with at most 8 unused bytes between files. Small file access is O(1) disk read.
SeaweedFS Filer uses off-the-shelf stores, such as MySql, Postgres, Sqlite, Mongodb, Redis, Elastic Search, Cassandra, HBase, MemSql, TiDB, CockroachCB, Etcd, YDB, to manage file directories. These stores are proven, scalable, and easier to manage.
| SeaweedFS | comparable to Ceph | advantage |
|---|---|---|
| Master | MDS | simpler |
| Volume | OSD | optimized for small files |
| Filer | Ceph FS | linearly scalable, Customizable, O(1) or O(logN) |
Compared to MinIO
Please note, as Apr 25, 2026 MinIO ceased development. It's strongly discouraged to use that unmaintained software with multiple security bugs.
MinIO followed AWS S3 closely and was ideal for testing for S3 API. It had good UI, policies, versionings, etc. SeaweedFS is trying to catch up here.
MinIO metadata were in simple files. Each file write will incur extra writes to corresponding meta file.
MinIO did not have optimization for lots of small files. The files were simply stored as is to local disks. Plus the extra meta file and shards for erasure coding, it only amplifies the LOSF problem.
MinIO had multiple disk IO to read one file. SeaweedFS has O(1) disk reads, even for erasure coded files.
MinIO had full-time erasure coding. SeaweedFS uses replication on hot data for faster speed and optionally applies erasure coding on warm data.
MinIO did not have POSIX-like API support.
MinIO had specific requirements on storage layout. It is not flexible to adjust capacity. In SeaweedFS, just start one volume server pointing to the master. That's all.
Dev Plan
- More tools and documentation, on how to manage and scale the system.
- Read and write stream data.
- Support structured data.
This is a super exciting project! And we need helpers and support!
Installation Guide
Installation guide for users who are not familiar with golang
Step 1: install go on your machine and setup the environment by following the instructions at:
https://golang.org/doc/install
make sure to define your $GOPATH
Step 2: checkout this repo:
git clone https://github.com/seaweedfs/seaweedfs.git
Step 3: download, compile, and install the project by executing the following command
cd seaweedfs/weed && make install
Once this is done, you will find the executable "weed" in your $GOPATH/bin directory
For more installation options, including how to run with Docker, see the Getting Started guide.
Disk Related Topics
Hard Drive Performance
When testing read performance on SeaweedFS, it basically becomes a performance test of your hard drive's random read speed. Hard drives usually get 100MB/s~200MB/s.
Solid State Disk
To modify or delete small files, SSD must delete a whole block at a time, and move content in existing blocks to a new block. SSD is fast when brand new, but will get fragmented over time and you have to garbage collect, compacting blocks. SeaweedFS is friendly to SSD since it is append-only. Deletion and compaction are done on volume level in the background, not slowing reading and not causing fragmentation.
Benchmark
My Own Unscientific Single Machine Results on Mac Book with Solid State Disk, CPU: 1 Intel Core i7 2.6GHz.
Write 1 million 1KB file:
Concurrency Level: 16
Time taken for tests: 66.753 seconds
Completed requests: 1048576
Failed requests: 0
Total transferred: 1106789009 bytes
Requests per second: 15708.23 [#/sec]
Transfer rate: 16191.69 [Kbytes/sec]
Connection Times (ms)
min avg max std
Total: 0.3 1.0 84.3 0.9
Percentage of the requests served within a certain time (ms)
50% 0.8 ms
66% 1.0 ms
75% 1.1 ms
80% 1.2 ms
90% 1.4 ms
95% 1.7 ms
98% 2.1 ms
99% 2.6 ms
100% 84.3 ms
Randomly read 1 million files:
Concurrency Level: 16
Time taken for tests: 22.301 seconds
Completed requests: 1048576
Failed requests: 0
Total transferred: 1106812873 bytes
Requests per second: 47019.38 [#/sec]
Transfer rate: 48467.57 [Kbytes/sec]
Connection Times (ms)
min avg max std
Total: 0.0 0.3 54.1 0.2
Percentage of the requests served within a certain time (ms)
50% 0.3 ms
90% 0.4 ms
98% 0.6 ms
99% 0.7 ms
100% 54.1 ms
Run WARP and launch a mixed benchmark.
make benchmark
warp: Benchmark data written to "warp-mixed-2025-12-05[194844]-kBpU.csv.zst"
Mixed operations.
Operation: DELETE, 10%, Concurrency: 20, Ran 42s.
* Throughput: 55.13 obj/s
Operation: GET, 45%, Concurrency: 20, Ran 42s.
* Throughput: 2477.45 MiB/s, 247.75 obj/s
Operation: PUT, 15%, Concurrency: 20, Ran 42s.
* Throughput: 825.85 MiB/s, 82.59 obj/s
Operation: STAT, 30%, Concurrency: 20, Ran 42s.
* Throughput: 165.27 obj/s
Cluster Total: 3302.88 MiB/s, 550.51 obj/s over 43s.
Enterprise
For enterprise users, please visit seaweedfs.com for the SeaweedFS Enterprise Edition, which has advanced features, including data recovery, self-healing storage, customizable erasure coding, EC vacuum and repair, etc.
License
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
The text of this page is available for modification and reuse under the terms of the Creative Commons Attribution-Sharealike 3.0 Unported License and the GNU Free Documentation License (unversioned, with no invariant sections, front-cover texts, or back-cover texts).
Stargazers over time
