diskmanager block device list should be sorted to avoid continual "block devices changed"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Fix Released
|
Medium
|
James Tunnicliffe | ||
1.25 |
Fix Released
|
High
|
Andrew Wilkins |
Bug Description
diskmanager detects that the block devices have constantly changed creating not only log spam, but most likely has an impact on database activity and size on the state server.
Most likely this list simply needs to be sorted consistently, see #1345832 for a previous similar issue with bootstrap server APIHostPort
Versions: client, state server and nodes all on xenial with 2.0-beta11-
root@semisomnol
02:43:46 up 15:19, 1 user, load average: 0.96, 1.28, 1.43
root@semisomnol
1671
Example log entries:
2016-07-10 02:35:58 INFO juju.worker.
2016-07-10 02:38:44 INFO juju.worker.
2016-07-10 02:40:29 INFO juju.worker.
Changed in juju-core: | |
status: | New → Triaged |
importance: | Undecided → Critical |
importance: | Critical → Medium |
Changed in juju-core: | |
status: | In Progress → Fix Committed |
Changed in juju-core: | |
status: | Fix Committed → Fix Released |
This is due to us sorting on the device name (e.g. sda), but then performing a deep comparison on everything we store, including the device links (by-id, by-path etc), which we don't sort.
Nice easy fix so I will get it queued up.