返回资讯中心
外部精选
软件工程
#开发者实时流

Ceph v21.2.0 Umbrella Released

Umbrella is the 21st stable release of Ceph.

Hacker News 最新taha_jahangir2026/10/31 · 17 天前30 分钟阅读

以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。

Umbrella is the 21st stable release of Ceph.

This is the first stable release of Ceph Umbrella. We recommend all users update to this release.

Warning: Umbrella container images are based on Rocky Linux 10 (el10), which requires a CPU supporting the x86-64-v3 microarchitecture level. Deployments on hosts with older CPUs will not be able to run the containerized daemons.

Warning: Python 3.9 or newer is required on all hosts for cephadm support.

Optimized ("Fast") Erasure Coding: a substantial new EC I/O path.

Ships with health warnings for deprecated EC plugins/techniques and for Blaum-Roth EC profiles whose w+1 isn't prime; non-4k-aligned chunk sizes are now rejected for optimized EC.

Security: CephX authentication hardening -- new AES256KRB5 cipher (replacing weaker legacy key types as the default), online key rotation for mon/OSD/MDS daemons during upgrades, and fixes for CVE-2025-30156 (AES-CBC misuse enabling authentication bypass), CVE-2026-39944, CVE-2026-50152, and CVE-2026-54330 (pr#71172).

Monitor: new optional RocksDB backup/restore mechanism (mon_backup_interval, ceph tell mon.* backup, offline restore via ceph-mon --restore-backup).

New availability score feature for Monitor/OSD (tracking + health warnings

Stretch clusters: tiebreaker mon can now be optional, added a max bucket-weight-diff threshold, a health warning for colocated monitors, and a restriction on changing the election strategy after stretch mode is enabled.

mgr/DaemonServer: ok-to-upgrade now aggregates and globally sorts OSDs across CRUSH child buckets (fixing inaccurate per-chassis/rack ordering), and upgrades can now be scoped to a CRUSH bucket.

pg-upmap-primary support and improved pg-upmap computing speed.

PG autoscaler: default mon_target_pg_per_osd increased 100 -> 200 and mon_max_pg_per_osd increased 250 -> 500, letting the autoscaler compute more parallelism-friendly pg_num values and improving the balancer's ability to reach uniform OSD utilization. Existing pools may autoscale to more PGs after upgrading as a result; pg_autoscale_during_upgrade (pr#67694) paces that splitting so it doesn't disrupt the upgrade itself. Clusters with a smaller-than-4GiB effective bluestore_osd_memory_target, or SSDs split into more than two OSDs, may want to pin the old values before upgrading (pr#69452).

BlueStore: BlueFS Spillover Cleaner evolution, faster allocation recovery, static/runtime object-fragmentation tracking, full-device discard on mkfs now default, several spanning-blob/onode-cache-stall/ reallocation-corruption fixes, and device-type-specific allocation policy.

mClock scheduler refactor. Default HDD OSD shard config changed as an interim fix for inconsistent client/recovery throughput with multiple shards: osd_op_num_shards_hdd 5 -> 1, osd_op_num_threads_per_shard_hdd 1 -> 5 (issue#66289).

Crimson/Seastore: continued heavy investment -- erasure coding support landed in Crimson (including recovery), PG splitting, a new 2Q cache replacement algorithm, broad coroutine conversions, and a long tail of correctness/perf fixes across the object-data, LBA/omap-btree, and cleaner subsystems.

(DRAFT) rgw/d4n: Add cache_request for cache API (pr#65314, Samarah Uriarte)

.github/milestone: add umbrella (pr#69407, Patrick Donnelly)

[DNM] ProtocolV2: make handle_existing_connection check for cookie mismatch first (pr#63413, Kamoltat Sirivadhna)

[Docs]: Corrected spelling mistake (pr#64088, Instelligence.io)

[mgr/dashboard]: Mismatch in the object size observed in Advance section in Tiering (pr#65170, Dnyaneshwari Talwekar)

[Packaging]: Use system-defined location for home directories (pr#62507, Shawn Edwards)

[run-make] Added flag for Dashboard and WError (pr#66181, Anshuman)

mgr/vol: handling the failed non-atomic operation (pr#59676, Neeraj Pratap Singh)

mds/MDSDaemon: unlock mds_lock while shutting down Beacon and others (pr#60326, Max Kellermann)

libcephfs: New feature - add ceph_setlk and ceph_getlk functions (pr#61116, Giorgos Kappes)

mds: for an irreparable hard link scrub is not able to identify the damage (pr#61173, Md Mahamudur Rahaman Sajib)

mds: add a proper check to evict clients with id 0 (pr#61981, neeraj pratap singh, Neeraj Pratap Singh)

mds: add ref counting to LogSegment (pr#62554, Milind Changire)

mds: Don't use global snaprealm's seq number for subvolume snapshots (pr#62682, Kotresh HR)

test/libcephfs: validate asynchronous write and fsync executing concurrently (issue#71515, pr#63636, Venky Shankar)

qa: Fix test_with_health_warn_with_2_active_MDSs (pr#64297, Kotresh HR)

mds: include auth credential in session dump (pr#64308, Patrick Donnelly)

client,mds,mgr: add subvolume level perf counters (pr#64434, Igor Golikov, Venky Shankar)

qa/suites/upgrade: add "Replacing daemon mds" to ignorelist (issue#50279, pr#64523, Venky Shankar)

client: crash caused by invalid iterator in _readdir_cache_cb (pr#64627, Zhansong Gao)

mds: Fix readdir when osd is full. (pr#64663, Kotresh HR)

mds: fix ineffective 'false' parameter due to virtual inheritance (pr#64702, izxl007)

Optimize the libcephfs proxy to reduce the overhead (pr#64716, Xavi Hernandez)

libcephfs_proxy: implement support for fscrypt (pr#64743, Xavi Hernandez)

mds: skip charmap handler check for MDS requests (pr#64822, Patrick Donnelly)

mds/MDSDaemon: Init lockup command millisecond timeout (pr#64958, Edwin Rodriguez)

mds/MDSMap: Enhance set_min_compat_client to validate iterator range (pr#64967, Edwin Rodriguez)

src/mds: Improve messages in MDCache.cc (pr#65232, Anthony D'Atri)

libcephfs_proxy: fix backward compatibility issue (pr#65320, Xavi Hernandez)

mds: uninline methods to reduce header dependencies and compile times (pr#65392, Max Kellermann)

mgr/volumes: fix bugs related missing UUID dir from a subvol (pr#65467, Rishabh Dave)

mds: fix rank 0 marked damaged if stopping fails after Elid flush. (pr#65483, ethanwu)

tools/cephfs: provide an estimate completion time for offline tools (pr#65495, Edwin Rodriguez)

libcephfs_proxy: add support for nonblocking fsync request (pr#65538, Xavi Hernandez)

cephfs-journal-tool: Journal trimming issue (pr#65577, Kotresh HR)

client: adjust Fb cap ref count check during synchronous fsync() (pr#65713, Venky Shankar)

mds: operator=() should return 'CDentry&' (pr#65876, Edwin Rodriguez)

mds: Simplify logging messages in quiesce_cluster_update (pr#65882, Edwin Rodriguez)

mds: Remove unnecessary std::move in MetricsHandler (pr#65884, Edwin Rodriguez)

mds: use after move ack error in QuiesceAgent::agent_thread_main (pr#65886, Edwin Rodriguez)

client: fix dump_mds_requests to valid json format (issue#73639, pr#66025, haoyixing)

test/libcephfs: Fixed path buffer overflow (pr#66030, Edwin Rodriguez)

client: signal waitfor_commit waiters for write delegation enabled inode (issue#73624, pr#66036, Venky Shankar)

mgr/volumes: remove unnecessary log error lines from earmark handling (pr#66093, Avan Thakkar)

libcephfs_proxy: implement batched readdir (pr#66183, Xavi Hernandez)

qa/cephfs: add ceph-osd-{classic,crimson} to exclude list of LTS releases (issue#74026, pr#66475, Venky Shankar)

libcephfs_proxy: Improvements to proxy configuration (pr#66907, Xavi Hernandez)

pybind/cephfs: invoke fcopyfile() libcephfs API without holding GIL (issue#74397, pr#66927, Venky Shankar)

qa: fix TypeError in delay (pr#67476, Jos Collin)

Add .clang-format configuration file (pr#65049, Edwin Rodriguez)

Add /health/snapshot api (pr#65019, Afreen Misbah)

Add Abhishek Desai into githubmap, organizationmap, mailmap (pr#66083, Abhishek Desai)

Add cephadm-signed certificate support for all services (pr#62106, Redouane Kachach)

add cephx update (pr#71172, Patrick Donnelly, Adam King, David Galloway, Sage McTaggart, Marcus Watts, Pritha Srivastava, Yehuda Sadeh, Matan Breizman)

Add Devika Babrekar into githubmap, organizationmap, mailmap (pr#66253, Devika Babrekar)

Add documentation for ceph_test_rados (pr#68437, Alex Ainscow)

Add health warning for deprecated EC plugins and techniques (pr#69026, Jamie Pryde)

Add missing includes to many subsystems (pr#64977, Max Kellermann)

Add normalization and casesensitive options to the subvolume group creation command (pr#64324, Venky Shankar, Xavi Hernandez)

Add NVMe preformat support to skip redundant mkfs discard (pr#66646, Ujjawal Anand)

Add PRE_BUILD to breakpad command (fix for CMake > 3.30) (pr#63643, Jesse F. Williamson)

add Sagar Gopale to githubmap mailmap organizationmap (pr#66436, Sagar Gopale)

Add Support for Catch2 unit tests (pr#62873, Jesse F. Williamson)

added code to fix failure on Haproxy error for rgw service with ipv6 (pr#67493, Kobi Ginon)

aio_cxx: Fix mutual deadlock and resolve test unreliability (pr#66430, Nitzan Mordechai)

Assorted build-with-container improvements (pr#63887, John Mulligan)

auth/cephx: harden service ticket verification (pr#72126, Dan van der Ster, Matthew Heler)

auth: add missing includes (pr#62844, Max Kellermann)

auth: msgr2 can return incorrect allowed_modes through AuthBadMethodFrame (pr#64717, Miki Patel)

auth: reduce header dependencies (pr#65541, Max Kellermann)

auth: remove unused AuthTicket::renew_after member variable (pr#64591, Kefu Chai)

Avoid calling zonegroup modify if realm/zonegroup/zone are not present (pr#67598, Kushal Deb)

backport nvmeof CNC and CLI teuthology tests (pr#72398, barakda, Vallari Agrawal)

Beacon diff (pr#65563, Leonid Chernin, Samuel Just)

blk,os/bluestore: do not call extblkdev::detect_device on every (pr#68177, Igor Fedotov)

blk/kernel : skip AIO thread for devices with size < block_size (pr#70380, Jaya Prakash)

blk/kernel: bring "bdev_async_discard" config parameter back (pr#65608, Igor Fedotov)

blk/kernel: improve DiscardThread life cycle (pr#64351, Igor Fedotov)

blk:Warning added for discard queue overflow (pr#61286, Jaya Prakash)

BlueFS: _flush_and_sync_log_jump_D should validate that jump_to >= log.writer->pos (pr#68043, dheart)

bluestore, extblkdev: Now plugins can raise health warnings (pr#68416, Adam Kupczyk)

bluestore/bluefs: FileWriter simpler flush (pr#66439, Igor Fedotov, Adam Kupczyk)

bluestore/BlueFS: fix bytes_written_slow counter with aio_write (pr#66147, chungfengz)

bluestore: Faster allocation recovery - evolution (pr#64369, Adam Kupczyk)

bluestore: Fix _setattr() with rare memory alignments (pr#63839, Adam Kupczyk, Alex Ainscow)

bluestore: Fix BlueFS WAL envelope mode rendering write buffer misalignment (pr#66438, Adam Kupczyk, Igor Fedotov)

bluestore: Fix unsharing blobs on remove (pr#70468, Adam Kupczyk)

Breakpad Crash Handler (pr#61264, Marcel Lauhoff)

btree: Safeguard delete_leaf_node to prevent dereferencing null nodes (pr#64961, Edwin Rodriguez)

build(deps): bump vite from 5.4.19 to 5.4.20 in /src/pybind/mgr/dashboard/frontend (pr#65571, dependabot[bot])

build-with-container: improve source rpm detection (pr#65156, John Mulligan)

build-with-container: add argument groups to organize options (pr#65514, John Mulligan)

build-with-container: build image variants (pr#65919, John Mulligan)

build-with-container: ensure npm dir is set up before configure (pr#65376, John Mulligan)

build-with-container: naming cleanup (pr#65835, John Mulligan)

build: Fix opentelemetry-cpp build failure on Noble (pr#62616, Adam C. Emerson)

build: Modify rocksdb submodule (pr#63190, Adam C. Emerson)

build: Rocky 9 support (pr#64657, David Galloway)

build: Update python packaging for src/python-common (pr#68401, John Mulligan)

Bump NFS Version in Ceph (pr#67906, Shweta Bhosale)

Carbonize upgrade page (pr#68971, Afreen Misbah)

ceph-backport.sh: clean function argument use and update GitHub help link (pr#67059, Ville Ojamo)

ceph-backport.sh: update known milestones (pr#63570, Jos Collin)

ceph-crash: reduce log noise from auth fallback in post_crash() (pr#68283, Kefu Chai)

ceph-exporter: fix systemd unit (pr#64561, Henry Hirsch)

ceph-mixin: Update monitoring mixin (pr#64755, Aashish Sharma, SuperQ, Ankush Behl

正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。

打开原文
去社区讨论 →