RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-28 14:34:13 +00:00
Code Issues Packages Projects Releases Wiki Activity
83,479 Commits 211 Branches 0 Tags 1.4 GiB
57ffaaa389
Commit Graph

83479 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Jason Dillaman
64c215e3dd
Merge pull request #20551 from shun-s/wip-add-notify-name-print 
rbd: output notifyOp request name when watching

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-27 07:47:05 -05:00
Radoslaw Zarzynski
360f823fac log: prohibit negative values for debug_* configurables. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:48 +01:00
Radoslaw Zarzynski
de5efc8170 log: remove SubsystemMap::m_max_name_len. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:48 +01:00
Radoslaw Zarzynski
d13e071d28 ceph: switch to SubsystemMap::should_gather<SubV, LvlV>(). 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:48 +01:00
Radoslaw Zarzynski
1a3e935774 core: resurrect the old check on dout's levels where possible. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:48 +01:00
Radoslaw Zarzynski
b265ed2955 core: hint the dout()'s message crafting as a cold code. 
The idea is to:
  1. Do not put the dout()'s crafting stuff on the hot, fall-through
     path. Cheapest branches are those that are forward and never taken.
  2. Move it to separated sections placed far away from the main path
     to be more friendly to ICache and ITLB. That is, dout_impl constructs
     a function now.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:48 +01:00
Radoslaw Zarzynski
d96bf57879 core: make SubsystemMap more statical and optimize should_gather(). 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:47 +01:00
Radoslaw Zarzynski
5b3c5bbf49 osd: remove the PGLog::prefix_provider as it's always nullptr. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:47 +01:00
Radoslaw Zarzynski
8a668355a1 common: move subsys-dependant types to a dedicated header. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:47 +01:00
Radoslaw Zarzynski
b0f94c628b core: dout checks ceph_subsys_* in compile time if possible. 
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
 2018-02-27 11:38:47 +01:00
Patrick Donnelly
c42015651e
Merge PR #19471 into master 
* refs/pull/19471/head:
	cephfs-journal-tool: support purge queue

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2018-02-27 02:22:45 -08:00
Patrick Donnelly
47ba679906
Merge PR #20412 into master 
* refs/pull/20412/head:
	mds/CDir: cleanup in the lookup_exact_snap()
	mds/cdir: add CDir missing state

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2018-02-27 02:22:38 -08:00
Patrick Donnelly
e016a6bc57
Merge PR #20468 into master 
* refs/pull/20468/head:
	mds: fix check of underwater dentries

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2018-02-27 02:22:31 -08:00
Patrick Donnelly
9c45cdff5f
Merge PR #20489 into master 
* refs/pull/20489/head:
	common: omit short option for id in help for clients

Reviewed-by: Jos Collin <jcollin@redhat.com>
 2018-02-27 02:22:25 -08:00
Patrick Donnelly
c66fcfae73
Merge PR #20547 into master 
* refs/pull/20547/head:
	mds: add uptime to MDS status

Reviewed-by: Douglas Fuller <dfuller@redhat.com>
 2018-02-27 02:22:19 -08:00
Patrick Donnelly
38cc4dbb82
Merge PR #20553 into master 
* refs/pull/20553/head:
	mds: always handle SESSION_REQUEST_RENEWCAPS messages
	mds: don't cleanup request that has pending remote authpin/wrlock/xlock

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2018-02-27 02:22:08 -08:00
lvshanchun
a55797e093 rgw: create sync-module instance when radosgw-admin sync run 
Signed-off-by: lvshanchun <lvshanchun@gmail.com>
 2018-02-27 16:14:08 +08:00
shangfufei
7705306465 msg/async:change EventCenter::process_events(int timeout_microseconds) to EventCenter::process_events(unsigned timeout_microseconds) 
Signed-off-by: shangfufei <shangfufei@inspur.com>
 2018-02-27 14:29:42 +08:00
Jason Dillaman
e6acc640c1
Merge pull request #20532 from Songweibin/wip-du-user 
rbd: filter out UserSnapshotNamespace in do_disk_usage

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 23:21:21 -05:00
Jason Dillaman
b3b3a400e1
Merge pull request #20529 from Songweibin/wip-fix-nbd-map 
rbd: set a default value for options in `nbd map`

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 23:20:59 -05:00
Jason Dillaman
698b6f8486
Merge pull request #20507 from trociny/wip-remote-rename 
librbd: make rename request complete with filtered code

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 23:20:24 -05:00
Haomai Wang
9765eb73da
Merge pull request #20600 from liewegas/wip-msgr-level 
msg/async/AsyncConnection: less noisy debug

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2018-02-27 10:05:29 +08:00
Gregory Farnum
6bd63de565
Merge pull request #20572 from batrick/erasure-code-cephfs-doc 
doc: cleanup erasure coded pool doc on cephfs use
 2018-02-26 14:59:59 -08:00
Greg Farnum
2d3fb03e19 bluestore: correctly check all block devices to decide if journal is_rotational 
Fixes: http://tracker.ceph.com/issues/23141

Signed-off-by: Greg Farnum <gfarnum@redhat.com>
 2018-02-26 13:36:07 -08:00
Sage Weil
3d07fa8bae msg/async/AsyncConnection: less noisy debug 
This spams logs of healthy OSDs.

Signed-off-by: Sage Weil <sage@redhat.com>
 2018-02-26 15:17:54 -06:00
Andrew Schoen
3d447dbf2b ceph-volume: lvm list now works with raw devices 
If you used a raw device with --data to create an OSD
you can now use 'lvm list' to get information about that OSD
using the raw device name like:

ceph-volume lvm list /dev/sda

fixes http://tracker.ceph.com/issues/23140

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
 2018-02-26 14:43:40 -06:00
David Zafman
a2a6f60d99
Merge pull request #20571 from dzafman/wip-21833-2 
Fix asserts caused by DNE pgs left behind after lots of OSD restarts

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
 2018-02-26 11:53:15 -08:00
Alfredo Deza
9c8afc2365
Merge pull request #20594 from ceph/wip-rm23004 
ceph-volume: log the current running command for easier debugging

Reviewed-by: Alfredo Deza <adeza@redhat.com>
 2018-02-26 14:24:38 -05:00
Jason Dillaman
c1b84297b8 qa/workunits/rbd: switch devstack tempest to 17.2.0 tag 
Fixes: http://tracker.ceph.com/issues/22961
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 14:04:48 -05:00
Mykola Golub
f2c329129b
Merge pull request #20560 from dillaman/wip-22960-1 
librbd: minor cleanup of the IO pathway

Reviewed-by: Mykola Golub <mgolub@suse.com>
 2018-02-26 20:51:36 +02:00
Jason Dillaman
18bce0a075 librbd: moved write same buffer assembler to common location 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 12:31:59 -05:00
Jason Dillaman
c2d47e7df1 librbd: decoupled object read requests from result assembler 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 12:31:59 -05:00
Jason Dillaman
ac077112d6 librbd: use move-semantics for object IO data payloads 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 12:31:58 -05:00
Jason Dillaman
49f4f331d9 librbd: track the originating source of a flush request 
This will be used in a later cleanup to remove the IO flush handling
from ImageCtx.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 12:29:37 -05:00
Jason Dillaman
baa0b68260 librbd: separated queued image IO requests from state machine 
This breaks the tight coupling between the IO work queue and
the actual dispatch of IO requests.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 12:28:25 -05:00
John Spray
2dca5b436a mgr: die on bind() failure 
Previously, the daemon would get wedged if it
competed for the same port as another daemon
on the same host and lost.

Fixes: https://tracker.ceph.com/issues/23037
Signed-off-by: John Spray <john.spray@redhat.com>
 2018-02-26 17:14:03 +00:00
Kefu Chai
92a2cbab46 ceph-objectstore-tool: do not check action_on_object_t::call()'s retval 
it never fails.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2018-02-27 01:08:23 +08:00
Kefu Chai
dd5d0e22f7 ceph-objectstore-tool: do not check queue_transaction()'s retval 
it never fails.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2018-02-27 01:08:22 +08:00
Kefu Chai
42f310f253 ceph-objectstore-tool: don't destroy SnapMapper until the txn is completed 
ch->flush() only ensure that the op is dequeued from osr/collection's
queue. but the ops could be still in the flight and they could still be
holding a *weak* reference of SnapMapper. we should make sure that the
mapper is still available when they reference it,

Fixes: http://tracker.ceph.com/issues/23121
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2018-02-27 01:08:22 +08:00
Andrew Schoen
911ca889cd ceph-volume: log the current running command for easier debugging 
Fixes: http://tracker.ceph.com/issues/23004

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
 2018-02-26 10:13:59 -06:00
David Zafman
5ca56072c3 osd: Remove partially created pg known as DNE 
Fixes: http://tracker.ceph.com/issues/21833

Signed-off-by: David Zafman <dzafman@redhat.com>
 2018-02-26 07:55:39 -08:00
Sage Weil
1378ab7df5
Merge pull request #20567 from liewegas/wip-kill-pg-preferred 
osd/osd_types: kill preferred field in pg_t

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
 2018-02-26 09:13:40 -06:00
Sage Weil
df455a8b9f osd/osd_types: pg_t: remove preferred 
We killed this many many years ago; remove it!

Keep the hash<pg_t> value the same just in case this leaked into some
on-disk or on-wire structure.

Signed-off-by: Sage Weil <sage@redhat.com>
 2018-02-26 09:06:57 -06:00
Alfredo Deza
91c877fb06
Merge pull request #20592 from ceph/ceph-volume-ssh 
ceph-volume: add ANSIBLE_SSH_RETRIES=5 to functional tests

Reviewed-by: Alfredo Deza <adeza@redhat.com>
 2018-02-26 09:55:55 -05:00
Piotr Dałek
136642da38 common/OpTracker: reorder fields 
Reorder smaller fields around so they're aligning naturally,
regaining a few bytes of storage.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
 2018-02-26 15:27:43 +01:00
Piotr Dałek
b4ba5e7c8e mon/Monitor: add missing shutdown of OpTracker 
Now that it has its own processing thread, it must be shut down
explicitly or it'll sigsegv randomly.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
 2018-02-26 15:27:43 +01:00
Piotr Dałek
6b75ee58b1 mon/OSDMonitor: remove op_tracker 
It's unused anyway.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
 2018-02-26 15:27:42 +01:00
Piotr Dałek
635a60d76f common/OpHistory: move insert/cleanup into separate thread 
Cluster that's flooded with incoming ops (and enabled optracker)
is bottlenecked by OpHistory::insert. Reduce that by:
- pushing incoming ops into separate queue that'll be processed by
  separate thread.
- using std::atomic_bool for shutdown flag so ops_history_lock doesn't
  need to be taken as often

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
 2018-02-26 15:27:42 +01:00
Ilya Dryomov
bdb7a75a6d
Merge pull request #20591 from idryomov/wip-krbd-latest-osdmap-on-map-test 
qa: krbd latest-osdmap-on-map test

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2018-02-26 15:07:01 +01:00
Andrew Schoen
44a5b2eb3a ceph-volume: add ANSIBLE_SSH_RETRIES=5 to functional tests 
Occasionally we get tests that fail because a test node becomes
'unreachable'. This should avoid those ssh connection issues we see sometimes
by increasing the amount of times ansible will try to reconnect to a node
after an ssh failure.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
 2018-02-26 07:54:56 -06:00