Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 455584 - sys-cluster/ceph-0.56.3 version bump
Summary: sys-cluster/ceph-0.56.3 version bump
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Server (show other bugs)
Hardware: All Linux
: Normal normal (vote)
Assignee: Gentoo Cluster Team
Depends on:
Reported: 2013-02-05 11:58 UTC by Tomáš Mózes
Modified: 2013-03-11 16:50 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Note You need to log in before you can comment on or make changes to this bug.
Description Tomáš Mózes 2013-02-05 11:58:01 UTC
v0.56.2 “bobtail”

    osd: snapshot trimming fixes
    osd: scrub snapshot metadata
    osd: fix osdmap trimming
    osd: misc peering fixes
    osd: stop heartbeating with peers if internal threads are stuck/hung
    osd: PG removal is friendlier to other workloads
    osd: fix recovery start delay (was causing very slow recovery)
    osd: fix scheduling of explicitly requested scrubs
    osd: fix scrub interval config options
    osd: improve recovery vs client io tuning
    osd: improve ‘slow request’ warning detail for better diagnosis
    osd: default CRUSH map now distributes across hosts, not OSDs
    osd: fix crash on 32-bit hosts triggered by librbd clients
    librbd: fix error handling when talking to older OSDs
    mon: fix a few rare crashes
    ceph command: ability to easily adjust CRUSH tunables
    radosgw: object copy does not copy source ACLs
    rados command: fix omap command usage
    sysvinit script: set ulimit -n properly on remote hosts
    msgr: fix narrow race with message queuing
    fixed compilation on some old distros (e.g., RHEL 5.x)

The ebuild for 0.56.1 seems to work for 0.56.2 (tested by using the volume from other machines).
Comment 1 Tomáš Mózes 2013-02-19 16:46:44 UTC
v0.56.3 “bobtail”

This release has several bug fixes surrounding OSD stability. Most significantly, an issue with OSDs being unresponsive shortly after startup (and occasionally crashing due to an internal heartbeat check) is resolved. Please upgrade.

    A bug was fixed in which the OSDMap epoch for PGs without any IO requests was not recorded. If there are pools in the cluster that are completely idle (for example, the data and metadata pools normally used by CephFS), and a large number of OSDMap epochs have elapsed since the ceph-osd daemon was last restarted, those maps will get reprocessed when the daemon restarts. This process can take a while if there are a lot of maps. A workaround is to ‘touch’ any idle pools with IO prior to restarting the daemons after packages are upgraded:

    rados bench 10 write -t 1 -b 4096 -p {POOLNAME}

    This will typically generate enough IO to touch every PG in the pool without generating significant cluster load, and also cleans up any temporary objects it creates.

Notable changes

    osd: flush peering work queue prior to start
    osd: persist osdmap epoch for idle PGs
    osd: fix and simplify connection handling for heartbeats
    osd: avoid crash on invalid admin command
    mon: fix rare races with monitor elections and commands
    mon: enforce that OSD reweights be between 0 and 1 (NOTE: not CRUSH weights)
    mon: approximate client, recovery bandwidth logging
    radosgw: fixed some XML formatting to conform to Swift API inconsistency
    radosgw: fix usage accounting bug; add repair tool
    radosgw: make fallback URI configurable (necessary on some web servers)
    librbd: fix handling for interrupted ‘unprotect’ operations
    mds, ceph-fuse: allow file and directory layouts to be modified via virtual xattrs