{"full_name":"cubefs/cubefs","owner":"cubefs","name":"cubefs","shard":"1","url":"https://github.com/cubefs/cubefs","description":"cloud-native distributed storage","homepage":"https://cubefs.io","topics":["distributed-storage","cncf","kubernetes","cloud-storage","erasure-coding","hybrid-cloud","ai-native-storage","object-storage","cloud-native-storage","distributed-file-system","data-orchestration","fuse"],"tag_pages":["cloud-storage","cncf","distributed-file-system","distributed-storage","fuse","kubernetes","object-storage"],"language":"Go","license":"Apache-2.0","avatar":"https://avatars.githubusercontent.com/u/33819338?v=4","created_at":"2019-02-19T03:15:04Z","pushed_at":"2026-11-13T07:56:22Z","metrics":{"stars":5697,"forks":716,"watchers":102,"issues":1109,"prs":2888,"releases":44,"issues_open":224,"issues_closed":885,"prs_open":47,"prs_closed":644,"prs_merged":2197,"discussions":110},"growth":{"day":{"stars":3,"forks":0,"base_day":"2026-11-15"}},"history":{"dates":["2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[5693,5692,5692,5692,5693,5694,5697],"forks":[715,715,716,716,716,716,716],"watchers":[null,102,102,102,102,102,102],"issues_open":[null,224,224,224,224,224,224],"prs_open":[null,47,47,47,47,47,47]},"boards":{},"same_owner":[],"issues":[{"number":1920,"title":"[Developer Activities]: 2023 The inaugural Developer Activities","comments":44,"created_at":"2023-04-18T01:47:21Z"},{"number":3650,"title":"[Bug]: performance tuning help need","comments":17,"created_at":"2025-01-19T09:09:43Z"},{"number":3173,"title":"[Bug]: the datanodes is not WRITABLE","comments":11,"created_at":"2024-03-05T03:27:56Z"},{"number":1580,"title":"[Bug Report] need add lock for snapshot newMetaItemIterator","comments":10,"created_at":"2022-10-27T06:04:23Z"},{"number":3922,"title":"[Enhancement]: Blobstore: add op stage latency timeout log for access/sdk/stream and blobnode","comments":8,"created_at":"2025-08-29T07:46:02Z"},{"number":3810,"title":"[Bug]: BlobStore: Read performance drops sharply from 18000 iops to 6000 iops","comments":8,"created_at":"2025-05-20T01:46:46Z"},{"number":3623,"title":"[Bug]: There is a high probability of bad file descriptor appearing in business reading.","comments":8,"created_at":"2025-01-06T07:24:51Z"},{"number":3064,"title":"[Feature]: 2024 plan","comments":8,"created_at":"2024-02-01T15:38:52Z"}],"discussions":[{"number":4118,"title":"how does the datanode handle the slow io and slow disk?","comments":1,"category":"Q&A","created_at":"2026-11-04T08:20:53Z"},{"number":4113,"title":"How can I change the install machine for master node?","comments":1,"category":"Q&A","created_at":"2026-10-16T08:02:51Z"},{"number":4110,"title":"where can I get the images source which can download by docker pull ?","comments":1,"category":"Q&A","created_at":"2026-10-12T12:50:32Z"},{"number":4097,"title":"rdma功能有排期计划吗?","comments":0,"category":"Q&A","created_at":"2026-09-08T09:33:28Z"},{"number":4096,"title":"how can I use one k8s node to deploy the cubefs ?","comments":1,"category":"Q&A","created_at":"2026-09-01T03:43:38Z"},{"number":4092,"title":"Title: [Bug/Question] software caused connection abort during large PVC restore with custom Go backup tool","comments":0,"category":"Q&A","created_at":"2026-08-24T06:13:35Z"},{"number":4088,"title":"how can I use the snapshot of the cubefs？","comments":1,"category":"Q&A","created_at":"2026-08-19T06:18:56Z"},{"number":4087,"title":"how can I do the stability test and durability test for cubefs?","comments":0,"category":"Q&A","created_at":"2026-08-14T11:29:42Z"}],"latest_release":{"name":"Release v3.6.0 - 2026/08/14","tag":"v3.6.0","published_at":"2026-08-12T09:21:44Z","body":"UPGRAGDE NOTICE\r\nIf you are using a CubeFS version earlier than v3.5.0, please refer to the UPGRADE NOTICE in version v3.5.0 for detailed upgrade steps and upgrade to v3.5.0 first.\r\n\r\nUpgrade nodes in this order: master → metanode → datanode → flashnode → objectnode → lcnode → cli → client.\r\n\r\n_Upgrade lcnode and flashnode when needed._\r\n\r\nBefore enabling MetaPartition learner-based decommissioning (`enableMpDecommissionByLearner`), upgrade all MetaNodes first.\r\n\r\nClients should use versions later than 3.2.0. Older versions need to be upgraded promptly; otherwise, there will be a risk of compromising stability.\r\n\r\n## Main Feature\r\n\r\n### Tag-based MP and DP migration ( #4072   @RyanWYR @aaronwu2010 )\r\n\r\n- Added tag-aware migration and decommissioning for MetaPartitions and DataPartitions.\r\n- Added per-tag concurrency limits and improved target-node selection, scheduling, and recovery control.\r\n\r\n### Multi-Region metadata support (#4071 @Victor1319 )\r\n\r\n- Added Region management for metadata nodes and volumes, including default Regions, allowed Region sets, and MetaPartition Region policies.\r\n- Added Region-aware placement validation, online policy updates, CLI operations, persistence, and monitoring metrics.\r\n\r\n### Multiple data storage pools ( #4076 @Victor1319 )\r\n\r\n- Improved DataPartition placement and accounting across multiple storage pools.\r\n- Added pool-aware writable DataPartition metrics and strengthened storage-mode, storage-type, and topology validation.\r\n- Supported lifecycle-based data rotation across storage pools.\r\n\r\n### Raft learners and learner-based MP decommissioning ( #4073 @Victor1319 )\r\n\r\n- Added Raft learner support to MetaPartition recovery and decommissioning workflows.\r\n- Improved learner selection, apply-progress validation, membership checks, and recovery scheduling to reduce service impact during replica replacement.\r\n\r\n### MP metadata alignment and self-healing ( #4074 @zhumingze1108 )\r\n\r\n- Added metadata alignment checks and automatic MetaPartition repair capabilities.\r\n- Improved recovery diagnostics, apply-progress gating, snapshot handling, and cross-version compatibility.\r\n\r\n### MetaNode proximity reads ( #4075 @Victor1319 )\r\n\r\n- Added Meta NearRead and follower-read support to route metadata reads to nearby, low-latency replicas.\r\n- Added follower-read leases and dirty-inode tracking to preserve read-your-writes consistency after metadata updates.\r\n\r\n### RocksDB experimental store mode ( #4083 @aaronwu2010 )\r\n\r\n- Added experimental RocksDB storage mode for MetaPartitions, with volume-level `defaultStoreMode` and per-replica `memory` / `rocksdb` selection.\r\n- Improved RocksDB disk selection, unavailable reporting on disk failure, and learner or follower placement that mixes memory and RocksDB store modes.\r\n\r\n## Enhancements\r\n\r\n- Optimized client read-ahead, retry backoff, batched directory reads, and inode/cache maintenance. ( #4077 @clinx  @bboyCH4 @leonrayang ). (\r\n- Improved extent migration with lease validation, inode generation updates, retry budgets, and clearer error classification. (#4081 @zhumingze1108 )\r\n- Improved FlashNode cache-key generation, capacity accounting, prewarm status handling, heartbeat processing, and management messages. ( #4082 @clinx @bboyCH4 )\r\n- Improved DataPartition decommission token scheduling, cancellation cleanup, progress handling, and automatic bad-disk decommissioning. ( #4079  @shuqiang-zheng )\r\n\r\n## Bug Fixes\r\n\r\n- Fixed stale client read-ahead data, cache-block double release, ticker leaks, and incomplete node-cache cleanup. (#4077 @bboyCH4 @clinx )\r\n- Fixed DataNode available-space underflow and delayed repair of new replicas with insufficient tiny extents. ( #4078 @shuqiang-zheng @RyanWYR )\r\n- Fixed DataPartition decommission token scheduling, cancellation cleanup, incorrect disk marking, and stalled bad-disk decommissioning. (#4078 @shuqiang-zheng )","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}