SeaweedFS
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!
- Download Binaries for different platforms
- SeaweedFS on Slack
- SeaweedFS Mailing List
- Wiki Documentation
- SeaweedFS Introduction Slides
Table of Contents
- Introduction
- Features
- Example Usage
- Architecture
- Compared to Other File Systems
- Dev Plan
- Installation Guide
- Disk Related Topics
- Benchmark
- License
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 an Object Store to handle small files efficiently. Instead of managing all file metadata in a central master, the central master only manages file volumes, and it lets these volume servers manage files and their metadata. This relieves concurrency pressure from the central master and spreads file metadata into volume servers, allowing faster file access (O(1), usually just one disk read operation).
SeaweedFS can transparently integrate with the cloud. With hot data on local cluster, and warm data on the cloud with O(1) access time, SeaweedFS can achieve both fast local access time and elastic cloud storage capacity, without any client side changes.
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
On top of the object store, optional Filer can support directories and POSIX attributes. Filer is a separate linearly-scalable stateless server with customizable metadata stores, e.g., MySql, Postgres, Redis, Etcd, Cassandra, LevelDB, MemSql, TiDB, TiKV, CockroachDB, etc.
Additional Features
- Can choose no replication or different replication levels, rack and data center aware.
- Automatic master servers failover - no single point of failure (SPOF).
- Automatic Gzip compression depending on file mime type.
- Automatic compaction to reclaim disk space after deletion or update.
- Servers in the same cluster can have different disk spaces, file systems, OS etc.
- Adding/Removing servers does not cause any data re-balancing.
- Optionally fix the orientation for jpeg pictures.
- 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.
- 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.
Filer Features
- filer server provide "normal" directories and files via http.
- mount filer to read and write files directly as a local directory via FUSE.
- Amazon S3 compatible API to access files with S3 tooling.
- Hadoop Compatible File System to access files from Hadoop/Spark/Flink/etc jobs.
- Async Backup To Cloud has extremely fast local access and backups to Amazon S3, Google Cloud Storage, Azure, BackBlaze.
- WebDAV access as a mapped drive on Mac and Windows, or from mobile devices.
Example Usage
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 -mserver="localhost:9333" -port=8080 &
> weed volume -dir="/tmp/data2" -max=10 -mserver="localhost:9333" -port=8081 &
Write File
To upload a file: 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 file 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 file 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 File 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 your own 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 File
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 file 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 File 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 file 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
Architecture
Usually distributed file systems split each file into chunks, a central master keeps a mapping of filenames, chunk indices to chunk handles, and also which chunks each chunk server has.
The main drawback is that the central master can't handle many small files efficiently, and since all read requests need to go through the chunk 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 files. 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 file metadata is stored in each volume on volume servers. Since each volume server only manages metadata of files on its own disk, with only 16 bytes for each file, all file access can read file 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 file. The client then contacts the volume node and POSTs the file content.
When a client needs to read a file 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.
Please see the example for details on the write-read process.
Storage Size
In the current implementation, each volume can hold 32 gibibytes (32GiB or 8x2^32 bytes). This is because we align content to 8 bytes. We can easily increase this to 64GiB, or 128GiB, or more, by changing 2 lines of code, at the cost of some wasted padding space due to alignment.
There can be 4 gibibytes (4GiB or 2^32 bytes) of volumes. So the total system size is 8 x 4GiB x 4GiB which is 128 exbibytes (128EiB or 2^67 bytes).
Each individual file size is limited to the volume size.
Saving memory
All file meta information stored on an 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.
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.
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 stores, e.g., Cassandra, Redis, Etcd, MySql, Postgres, MemSql, TiDB, CockroachDB, etc, and is easy to customized.
- SeaweedFS Volume server also communicates directly with clients via HTTP, supporting range queries, direct uploads, etc.
