Actions
Bug #20922
openmisdirected op with localize_reads set
Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
2017-08-05 04:49:50.555762 7ff660852700 10 osd.1 155 dequeue_op 0x555dc8fe32c0 prio 63 cost 0 latency 8.137564 osd_op(client.4192.0:200970 3.7 3.c22d258f (undecoded) ondisk+retry+read+localize_reads+known_if_redirected e153) v8 pg pg[3.7( v 148'2326 (78'826,148'2326] lb MIN (bitwise) local-lis/les=154/155 n=0 ec=25/25 lis/c 108/108 les/c/f 109/109/0 153/154/108) [3,1]/[3,4] r=-1 lpr=155 pi=[108,154)/1 crt=148'2326 lcod 0'0 unknown] 2017-08-05 04:49:50.555769 7ff661053700 0 log_channel(cluster) log [WRN] : client.4192 172.21.15.203:0/851039524 misdirected client.4192.0:200864 pg 3.a4ed2e86 to osd.1 not [3,4] in e153/155
(several instances of this)
/a/sage-2017-08-04_23:16:09-rbd-wip-sage-testing2-20170804b-distro-basic-smithi/1484059
rbd/thrash/{base/install.yaml clusters/{fixed-2.yaml openstack.yaml} msgr-failures/few.yaml objectstore/bluestore.yaml thrashers/cache.yaml thrashosds-health.yaml workloads/rbd_fsx_cache_writethrough.yaml}
Updated by Greg Farnum over 6 years ago
Well, the issue is not immediately apparent, but _calc_target() is pretty complicated and we're feeding in a not-totally-reset op_target_t structure when ops get retried on EAGAIN, so it wouldn't surprise me if there was some issue there.
Updated by Greg Farnum over 6 years ago
- Priority changed from High to Normal
Downgrading this as it's localized_reads.
Actions