]>
git.apps.os.sepia.ceph.com Git - ceph.git/log
myoungwon oh [Tue, 10 Oct 2017 14:16:26 +0000 (23:16 +0900)]
osd: use get_pgid() instead of info.pgid
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
myoungwon oh [Fri, 22 Sep 2017 07:10:01 +0000 (16:10 +0900)]
osd: use global id instead of entity_inst_t
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
Kefu Chai [Wed, 12 Jul 2017 07:19:13 +0000 (15:19 +0800)]
osd: move pgid into OpQueueItem
this pave the road to generalize OpWQ
We're going to want to be able to queue things that are not ordered by
the PG lock. To that end, this patch genearlizes OSD::ShardedOpWQ to
use a type which can specify an ordering token and locking structure
other than a PG.
There is a lot of collateral damage which I didn't feel was worth
separating out into other commits. The code in ShardedOpWQ itself got
some superficial cleanup. Also, the item being queued has been switched
to not use a boost::variant. It was a cute way before to make the type
easily copyable, but adding more visitors for the locking support would
have been annoying. Instead, the variant is a unique_ptr to an
interface. This makes the queue item type no longer copyable, which is
just as well since we don't really want to be copying queue items anyway
(duplicates would be most likely a bug)
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
myoungwon oh [Fri, 15 Sep 2017 16:04:43 +0000 (01:04 +0900)]
src/test: fix move semantics
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
Kefu Chai [Wed, 12 Jul 2017 04:11:45 +0000 (12:11 +0800)]
common,osd: enforce move semantics in WorkQueue
enable the WorkQueue to support move-only template types.
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
Kefu Chai [Wed, 12 Jul 2017 02:28:37 +0000 (10:28 +0800)]
enforce move semantics of Request type
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
Kefu Chai [Fri, 30 Jun 2017 07:13:48 +0000 (15:13 +0800)]
OSD: s/PGQueueable/OpQueueItem/g
PGQueueable was always a pretty bad name, and going forward, we won't
necessarily be scheduling things ordered only by the PG lock, so let's
change it.
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
Samuel Just [Tue, 31 May 2016 22:28:24 +0000 (15:28 -0700)]
PG.h: remove some unnecessary virtual declarations
PG users do not need these public methods.
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
Haomai Wang [Mon, 9 Oct 2017 13:48:19 +0000 (08:48 -0500)]
Merge pull request #18190 from JustL/master
msg/async/rdma: unnecessary reinitiliazation of an iterator
Reviewed-by: Haomai Wang <haomai@xsky.com>
Sage Weil [Mon, 9 Oct 2017 13:22:42 +0000 (08:22 -0500)]
Merge pull request #18058 from Yan-waller/wip-walle-0927pgstatexpandstate
osd/PG: extend pg state bits to fix pg ls commands error
Reviewed-by: Sage Weil <sage@redhat.com>
JustL [Mon, 9 Oct 2017 12:37:43 +0000 (20:37 +0800)]
msg/async/rdma: unnecessary reinitiliazation of an iterator
current_buffer iterator does not need to be reinitalized as no changes
occur between the initialization and iteration steps.
Signed-off-by: Justinas Lingys jlingys@connect.ust.hk
Orit Wasserman [Mon, 9 Oct 2017 12:19:05 +0000 (15:19 +0300)]
Merge pull request #18178 from joscollin/wip-radosgw-admin-reshard-fix
rgw: update usage() with status
Reviewed-by: Orit Wasserman <owasserm@redhat.com>
Jos Collin [Mon, 9 Oct 2017 12:03:57 +0000 (12:03 +0000)]
Merge pull request #18180 from oritwas/wip-rgw-21617
rgw: add reshard commands
Reviewed-by: Jos Collin <jcollin@redhat.com>
Jos Collin [Mon, 9 Oct 2017 11:28:29 +0000 (16:58 +0530)]
rgw: Fixed typos and status in help.t
Fixed typos and status in help.t
Signed-off-by: Jos Collin <jcollin@redhat.com>
Orit Wasserman [Mon, 9 Oct 2017 09:35:42 +0000 (12:35 +0300)]
rgw: add reshard commands to the man pages
Fixes: http://tracker.ceph.com/issues/21617
Signed-off-by: Orit Wasserman <owasserm@redhat.com>
Kefu Chai [Mon, 9 Oct 2017 10:25:45 +0000 (18:25 +0800)]
Merge pull request #18162 from tchaikov/wip-ceph-disk-retry-on-dev-not-ready
ceph-disk: retry on OSError
Reviewed-by: Loic Dachary <ldachary@redhat.com>
Jos Collin [Mon, 9 Oct 2017 09:14:29 +0000 (14:44 +0530)]
rgw: update usage() with status
Fixed the missing 'status' in usage().
Signed-off-by: Jos Collin <jcollin@redhat.com>
Yan Jun [Wed, 27 Sep 2017 08:35:14 +0000 (16:35 +0800)]
osd/PG: extend pg state to 64 bits
Signed-off-by: Yan Jun <yan.jun8@zte.com.cn>
Yan Jun [Wed, 27 Sep 2017 11:02:36 +0000 (19:02 +0800)]
os/bluestore: drop unused class FreelistManager
Signed-off-by: Yan Jun <yan.jun8@zte.com.cn>
Jos Collin [Mon, 9 Oct 2017 03:27:27 +0000 (03:27 +0000)]
Merge pull request #18157 from batrick/ptl-tool-branch-checkout
ptl-tool: checkout branch after creation
Reviewed-by: Jos Collin <jcollin@redhat.com>
Sage Weil [Mon, 9 Oct 2017 02:47:49 +0000 (21:47 -0500)]
Merge pull request #17708 from liewegas/wip-pg
osd: initial minimal efforts to clean up PG interface
Sage Weil [Mon, 9 Oct 2017 02:47:26 +0000 (21:47 -0500)]
Merge pull request #18125 from cernceph/dvanders_conf
mon: mark mon_allow_pool_delete as observed
Reviewed-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 9 Oct 2017 02:47:02 +0000 (21:47 -0500)]
Merge pull request #18144 from hjwsm1989/wip-21425-part2
qa/suites/rest/basic/tasks/rest_test: whiltelist OSD_DOWN
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Sage Weil [Mon, 9 Oct 2017 02:46:27 +0000 (21:46 -0500)]
Merge pull request #18159 from liewegas/wip-21712-b
os: fix 0-length zero semantics, add tests
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Sage Weil [Fri, 6 Oct 2017 20:29:32 +0000 (15:29 -0500)]
os/bluestore: 0-length zero should not change object size
Fixes: http://tracker.ceph.com/issues/21712
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 20:29:15 +0000 (15:29 -0500)]
os/filestore: make 0-length zero avoid touching object length
Fixes: http://tracker.ceph.com/issues/21712
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 20:28:49 +0000 (15:28 -0500)]
os/ObjectStore: 0-length zero does not change length of object
Signed-off-by: Sage Weil <sage@redhat.com>
Kefu Chai [Mon, 9 Oct 2017 01:35:46 +0000 (09:35 +0800)]
Merge pull request #18141 from adamemerson/wip-log-time
log: Use the coarse real time clock in log timestamps
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Kefu Chai [Mon, 9 Oct 2017 01:34:50 +0000 (09:34 +0800)]
Merge pull request #17976 from adamemerson/wip-warning-sign
mgr,osd,kv: Fix various warnings for Clang and GCC7
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Kefu Chai [Sat, 7 Oct 2017 14:15:11 +0000 (22:15 +0800)]
ceph-disk: retry on OSError
we are likely to
1) create partition, for instance, sdc1
2) partprobe sdc
3) udevadm settle
4) check the device by its path: /dev/sdc1
but there is chance that the uevent sent from kernel fails to reach udev
before we call "udevadm", hence "/dev/sdc1" does not exist even after
"udevadm settle" returns. so we retry in case of OSError here.
Signed-off-by: Kefu Chai <kchai@redhat.com>
Kefu Chai [Sat, 7 Oct 2017 13:13:47 +0000 (21:13 +0800)]
ceph-disk: factor out the retry logic into a decorator
Signed-off-by: Kefu Chai <kchai@redhat.com>
Adam C. Emerson [Fri, 6 Oct 2017 21:15:43 +0000 (17:15 -0400)]
log: Mark timestamps for coarseness and print appropriately
Have each timestamp know whether it's coarse or fine. Print six figures
for fine stamps and three for coarse.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Adam C. Emerson [Thu, 5 Oct 2017 22:34:06 +0000 (18:34 -0400)]
log: Add log_coarse_timestamps option
And set up an observer in CephContext to coarsen/refine time stamps as
requested.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Adam C. Emerson [Thu, 5 Oct 2017 22:32:32 +0000 (18:32 -0400)]
config: Add templated _get_val for use in observers
When notified by a config observer, the config lock is already held,
create a _get_val template to manipulate newstyle options when
notified.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Adam C. Emerson [Thu, 5 Oct 2017 22:30:56 +0000 (18:30 -0400)]
log: Use log_clock, defaulting to coarse time
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Adam C. Emerson [Thu, 5 Oct 2017 22:27:37 +0000 (18:27 -0400)]
log: Add optional msg parameter to create_entry and test with it
The tests should, really, get their time from the same source as the
log, and also probably use the same interface to creating entries that
other people do. So add an optional message argument and have the
tests use that rather than using the new operator directly.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Adam C. Emerson [Thu, 5 Oct 2017 22:17:16 +0000 (18:17 -0400)]
log: Create log_clock to replace use of utime_t
We create a new time type to support using coarse timestamps (for
performance) by default while still allowing the user to switch to
fine timestamps.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Sage Weil [Fri, 6 Oct 2017 19:34:41 +0000 (14:34 -0500)]
osd/PGLog: drop old compat coll/oid args to read_log_and_missing
The oldest version we care about is 10, and these were for <8.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 20:32:45 +0000 (15:32 -0500)]
Merge pull request #18122 from liewegas/wip-monthrash-peering
qa/suites/rados/monthrash: tolerate PG_AVAILABILITY during mon thrashing
Yuri Weinstein [Fri, 6 Oct 2017 19:48:35 +0000 (12:48 -0700)]
Merge pull request #18047 from gregsfortytwo/wip-20416-bitwise-assert
osd: make the PG's SORTBITWISE assert a more generous shutdown
Reviewed-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 19:21:43 +0000 (14:21 -0500)]
ceph-objectstore-tool: bring past_intervals up to date on import
We can't blinding fast-forward the pg epoch to the latest map without
also updating the pg history and PastIntervals. Do that at import time.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 18:58:22 +0000 (13:58 -0500)]
ceph-objectstore-tool: hide DIAGNOSTIC stuff behind debug/verbose mode
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 18:57:59 +0000 (13:57 -0500)]
ceph-objectstore-tool: drop unused biginfo_oid
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 6 Oct 2017 18:43:57 +0000 (13:43 -0500)]
ceph-objectstore-tool: refuse to import PG older that OSD's oldest_epoch
We don't have a way to construct a valid PastIntervals history.
Signed-off-by: Sage Weil <sage@redhat.com>
Patrick Donnelly [Fri, 6 Oct 2017 18:52:04 +0000 (11:52 -0700)]
ptl-tool: checkout branch after creation
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
Sage Weil [Thu, 5 Oct 2017 02:31:12 +0000 (21:31 -0500)]
ceph-objectstore-tool: remove rm-past-intervals op
The OSD doesn't rebuild this on demand anymore.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Tue, 3 Oct 2017 18:32:37 +0000 (13:32 -0500)]
ceph-objectstore-tool: do not clear same_interval_since and PastIntervals
We dropped the OSD recalculation code for this. Import it.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 23:27:44 +0000 (18:27 -0500)]
ceph-objectstore-tool: print pg_info_t to stdoutt
Helpful for debugging
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 21:53:31 +0000 (16:53 -0500)]
osd: print summary for forced backfill/recovery to debug
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:53:47 +0000 (14:53 -0500)]
osd/PG: drop OSD friend, finally!
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:51:36 +0000 (14:51 -0500)]
osd: drop redundant is_active() check before scrub_sched()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:14:11 +0000 (14:14 -0500)]
osd/PG: make update_snap_mapper_bits public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:12:08 +0000 (14:12 -0500)]
osd/PG: make on_removal() public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:11:49 +0000 (14:11 -0500)]
osd/PG: make get_primary() public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:05:00 +0000 (14:05 -0500)]
osd/PG: add get_pool()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:04:44 +0000 (14:04 -0500)]
osd: use pg_id.pool() for pg pool id
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:04:14 +0000 (14:04 -0500)]
osd/PG: make sched_scrub() public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:01:00 +0000 (14:01 -0500)]
osd/PG: get_must_scrub()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 19:00:37 +0000 (14:00 -0500)]
osd/PGBackend: remove scrub_supported()
Everybody supports scrub.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 18:59:33 +0000 (13:59 -0500)]
osd/PG: change 'debug dump_missing'
- json
- move into PG class
No users in the qa suite.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 02:55:27 +0000 (21:55 -0500)]
osd/PG: move bits of delete_dir (pg removal) into PG
We don't want to expose things like SnapMapper, OSDriver.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 02:32:44 +0000 (21:32 -0500)]
osd/PG: define public interface for heartbeat peers
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Mon, 18 Sep 2017 02:17:36 +0000 (21:17 -0500)]
osd: use accessors for pg resurrection
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 23:32:45 +0000 (18:32 -0500)]
osd/PG: simplify start_recovery_ops return value meaning
Instead of complex check in caller, return simple bool indicating whether
find_unfound() should be called with a rctx.
There is one remaining condition that can probably be simplified:
if (!recovering.empty() ||
work_in_progress || recovery_ops_active > 0 || deferred_backfill)
return !work_in_progress && have_unfound();
but we will leave it for now.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 23:29:38 +0000 (18:29 -0500)]
osd/PG: simplify unfound recovery case; rename method
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 23:27:44 +0000 (18:27 -0500)]
osd/PG: drop unneeded write_if_dirty
Nothing in start_recovery_ops or stuck_on_unfound dirties state.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 23:11:35 +0000 (18:11 -0500)]
osd/PG: move a few assertions, flag change into start_recovery_ops()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:55:29 +0000 (17:55 -0500)]
osd/PG: move some more split machinery into PG
and move other split pieces into the public interface.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:37:36 +0000 (17:37 -0500)]
osd/PG: move write_if_dirty inside handle_{loaded,create}
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:35:31 +0000 (17:35 -0500)]
osd/PG: get_need_up_thru(), get_same_interval_since()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:30:32 +0000 (17:30 -0500)]
osd: bit less verbose while loading pgs
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:29:16 +0000 (17:29 -0500)]
osd/PG: move more recovery logic into PG
I suspect we eventually want to move the create_context and
dispatch_context into OSDService (if it isn't there already) and move
even more of this logic into PG.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:10:53 +0000 (17:10 -0500)]
osd/PG: move some osd logic into handle_pg_trim
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sun, 17 Sep 2017 22:02:05 +0000 (17:02 -0500)]
osd/PG: hide peering_queue from OSD
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 18:08:07 +0000 (14:08 -0400)]
osd: remove build_past_intervals_parallel
This was added in the dumpling era to handle the case where we have lots
of PGs with stale or missing past_intervals. These days, we update
PastIntervals as we inject new makes (via advance_map), when we
learn about new PGs via notify with get the PastIntervals at the same time,
and during PGCreate handling we have a separate path in
build_initial_pg_history. This code is obsolete.
It also reaches deep into the PG class where it shouldn't!
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:53:57 +0000 (23:53 -0400)]
osd: use PG::is_deleting()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:52:33 +0000 (23:52 -0400)]
osd/PG: move load_pgs cruft into read_state()
This only happens during initial load; move it there.
Also, since update_store_with_options
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:44:53 +0000 (23:44 -0400)]
osd/PG: make read_state, peek_map_epoch public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:31:54 +0000 (23:31 -0400)]
osd/PG: simplify/streamline upgrade process
There's no reason for this to be exposed to the OSD. Just do it directly
in read_state().
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:26:17 +0000 (23:26 -0400)]
osd/PG: drop bufferlist passed from peek_map_epoch to read_state/read_info
This is no longer used or needed.
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:15:19 +0000 (23:15 -0400)]
osd/PG: make init() public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:11:34 +0000 (23:11 -0400)]
osd/PG: on_shutdown -> shutdown in public interface
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:08:33 +0000 (23:08 -0400)]
osd/PG: add get_num_ref()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 03:07:38 +0000 (23:07 -0400)]
osd/PG: add get_pg_stats() to public interface
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 02:58:15 +0000 (22:58 -0400)]
osd/PG: add dump_pgstate_history() to public interface
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Sat, 16 Sep 2017 02:53:20 +0000 (22:53 -0400)]
osd/PG: make get_watchers public, hide locking
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:43:29 +0000 (18:43 -0400)]
osd: drop legacy infos object cleanup
This won't be there post-luminous
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:41:35 +0000 (18:41 -0400)]
osd/PG: make osr public
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:41:09 +0000 (18:41 -0400)]
osd: use is_deleting()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:41:03 +0000 (18:41 -0400)]
osd: use pg->pg_id const member, not get_pgid()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:35:40 +0000 (18:35 -0400)]
osd/PG: reorder some members
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:22:01 +0000 (18:22 -0400)]
osd/PG: protect get_last_peering_reset
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 22:13:19 +0000 (18:13 -0400)]
osd/PG: protect get_role()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 15:37:29 +0000 (11:37 -0400)]
osd/PG: cleanup and move ref counting public bits
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 15:26:16 +0000 (11:26 -0400)]
osd/PG: protect pg state accessors
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 15:24:45 +0000 (11:24 -0400)]
osd/PG: add is_forced_recovery_or_backfill()
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 15:22:53 +0000 (11:22 -0400)]
osd/ECBackend: use Listener instead of PG interface to check undersized
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Fri, 15 Sep 2017 15:17:48 +0000 (11:17 -0400)]
osd/PG: collection more of public at top of class
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Thu, 14 Sep 2017 13:44:06 +0000 (09:44 -0400)]
osd/PG: expose static methods for ceph-objectstore-tool
Signed-off-by: Sage Weil <sage@redhat.com>
Sage Weil [Wed, 13 Sep 2017 22:14:59 +0000 (18:14 -0400)]
osd/PG: move C_PG_FinishRecovery into header
Less code than delcaring it in 2 places (by making it a friend).
Signed-off-by: Sage Weil <sage@redhat.com>