Swift replicator dies but leaves behind pidfile
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Swift Storage Charm |
Invalid
|
Undecided
|
Unassigned | ||
swift (Ubuntu) |
Triaged
|
Medium
|
Unassigned |
Bug Description
Seeing this on 2 Mitaka/Trusty clouds at the moment. Swift is 2.7.1 and 2.7.0.
Juju 1.25.10 with 16.10 charms
Juju 1.25.6 with 17.02 charms
Replicators frequently fall over. Sometimes we can map these to hooks running and other times not. The main problem we see is that a pidfile is left behind so the service can not be automatically restarted without manual intervention to issue a hard stop first, or rm'ing the pidfile.
root@mycloud-
root 4556 0.0 0.0 10480 2072 pts/3 S+ 00:20 0:00 \_ grep --color=auto replicator
swift 8586 0.1 0.0 84848 27364 ? Ss Apr05 93:13 /usr/bin/python /usr/bin/
swift 14251 0.2 0.1 1582612 45848 ? Ssl Apr05 147:59 /usr/bin/python /usr/bin/
root@mycloud-
account-auditor.pid account-
account-reaper.pid account-server.pid container-
root@mycloud-
1352
root@mycloud-
strace: attach: ptrace(
root@mycloud-
swift-container
root@mycloud-
swift-container
root@mycloud-
root@mycloud-
stop: Job failed while stopping
root@mycloud-
swift-container
root@mycloud-
root 29776 0.0 0.0 10480 2068 pts/3 S+ 00:59 0:00 \_ grep --color=auto replicator
swift 8586 0.1 0.0 84848 27364 ? Ss Apr05 93:16 /usr/bin/python /usr/bin/
swift 14251 0.2 0.1 1582612 45848 ? Ssl Apr05 148:04 /usr/bin/python /usr/bin/
swift 26898 0.1 0.0 98708 30600 ? Ss 00:35 0:02 /usr/bin/python /usr/bin/
This continues to be a recurrent problem. Is there any additional information we can provide that would help?