We always do a restart on config change if we have persistent storage, even if we've set it to reload on config change
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
postgresql (Juju Charms Collection) |
Fix Released
|
Undecided
|
Stuart Bishop |
Bug Description
<mthaddon> so it seems there's something up with the DB shutting down intermittently
<mthaddon> https:/
<mthaddon> trying to figure out what's going on there
<mthaddon> 2013-05-02 09:42:27,470: unit.hook.api@INFO: about reconfigure service with config_
<gnuoy> mthaddon, I see entries in the charm.log on the pg server which tie in
<mthaddon> :)
<mthaddon> gnuoy: seems like we need to change that config value to do reloads
<gnuoy> yeah!
<mthaddon> that's weird, the default is reload
<gnuoy> mthaddon, shall I do that or are you doing it?
<mthaddon> gnuoy: taking a look now, will let you know before I change anything
<mthaddon> juju get cert-postgresql suggests it should be doing a reload, not a restart
<bigkevmcd> seems to be completing tasks again
<mthaddon> so I think this is a charm bug
<mthaddon> we do "if config_
<mthaddon> so we are basically always going to do a restart if we have a persistent storage volume
Related branches
- Tom Haddon: Approve
-
Diff: 11 lines (+0/-1)1 file modifiedhooks/hooks.py (+0/-1)
tags: | added: canonical-webops |
Changed in postgresql (Juju Charms Collection): | |
status: | In Progress → Fix Released |
config_ change_ command is no longer wired up to anything.
Restarting is happening automatically whenever postgresql.conf entries are changed that require it.
Other cases get a reload. Except for bits like this one, which IIRC already had a forced restart and I didn't want to change it.