Bug #2164
osd: scrub missing _, snapset attrs
0%
Description
2012-03-11 05:43:14.226733 osd.14 10.214.133.31:6800/8629 41 : [ERR] 0.32 osd.11: soid 2f105832/plana15_2169_object7344/head ex
tra attr _, extra attr snapset
2012-03-11 05:43:14.226761 osd.14 10.214.133.31:6800/8629 42 : [ERR] 0.32 scrub 0 missing, 1 inconsistent objects
2012-03-11 05:43:14.227430 osd.14 10.214.133.31:6800/8629 43 : [ERR] scrub 0.32 2f105832/plana15_2169_object7344/head no 'snapset' attr
2012-03-11 05:43:14.227526 osd.14 10.214.133.31:6800/8629 44 : [ERR] 0.32 scrub stat mismatch, got 210/211 objects, 0/0 clones, 880803840/884998144 bytes.
2012-03-11 05:43:14.227543 osd.14 10.214.133.31:6800/8629 45 : [ERR] 0.32 scrub 2 errors
on
ubuntu@teuthology:/a/master-2012-03-10_21:22:04/413
kernel: &id001
branch: testing
nuke-on-error: true
overrides:
ceph:
branch: master
btrfs: 1
log-whitelist:
- clocks not synchronized
- old request
roles:
- - mon.0
- mds.a
- osd.0
- - mon.1
- osd.1
- - mon.2
- osd.2
- - osd.3
- - osd.4
- - osd.5
- - osd.6
- - osd.7
- - osd.8
- - osd.9
- - osd.10
- - osd.11
- - osd.12
- - osd.13
- - osd.14
- - osd.15
- - client.0
tasks:
- internal.lock_machines: 17
- internal.save_config: null
- internal.check_lock: null
- internal.connect: null
- internal.check_conflict: null
- kernel: *id001
- internal.base: null
- internal.archive: null
- internal.coredump: null
- internal.syslog: null
- internal.timer: null
- ceph:
log-whitelist:
- wrongly marked me down or wrong addr
- objects unfound and apparently lost
- thrashosds:
chance_down: 50
- radosbench:
clients:
- client.0
time: 1800
Related issues
Associated revisions
filestore: set guard on collection_move
During recovery we submit transactions like:
- delete a/foo
- move tmp/foo to a/fooThis prevents the EEXIST check in collection_move from doing any good,
since the destination never exists. We need to do that remove at least
sometimes, because we may be overwriting an existing/older version of the
object.
So,
- set the guard after we do the move, so that
- the delete won't be repated, and
- the EEXIST check will work
Also check the guard for good measure (although that doesn't do anything
specifically useful in this scenario).
Fixes: #2164
Signed-off-by: Sage Weil <sage@newdream.net>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Reviewed-by: Samuel Just <samuel.just@dreamhost.com>
History
#1 Updated by Sage Weil over 11 years ago
- Status changed from New to 12
this was non-btrfs, right after the new idempotent replay stuff was fixed.
#2 Updated by Sage Weil over 11 years ago
- Status changed from 12 to 7
- Target version set to v0.45
wip-2164
it's a problem with the collection_move guard (or lack thereof)
#3 Updated by Sage Weil over 11 years ago
- Status changed from 7 to Resolved