Galera lxc shutdown results in corrupted db
Bug #1806696 reported by
Justin Alford
This bug affects 3 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack-Ansible |
Fix Released
|
Undecided
|
Justin Alford |
Bug Description
After a recent set of controller reboots, our Galera cluster didn't come back up and had to be rebuilt. This appears to be due to the fact that the lxc service is configured to forcibly kill containers after 5 seconds if they aren't down by then (due to the SHUTDOWNDELAY env var being set to 5).
summary: |
- Galera lxc shutdown + Galera lxc shutdown results in corrupted db |
Changed in openstack-ansible: | |
status: | New → Confirmed |
Changed in openstack-ansible: | |
assignee: | nobody → Justin Alford (jlalford) |
Changed in openstack-ansible: | |
status: | Confirmed → Fix Released |
To post a comment you must log in.
root@controller -dc1r02n03: ~# systemctl show lxc.service |grep ExecStop lib/x86_ 64-linux- gnu/lxc/ lxc-containers ; argv[]= /usr/lib/ x86_64- linux-gnu/ lxc/lxc- containers stop ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }
ExecStop={ path=/usr/
/usr/lib/ x86_64- linux-gnu/ lxc/lxc- containers:
SHUTDOWNDELAY=5 AY="-t $SHUTDOWNDELAY" /lxc-autostart $STOPOPTS $SHUTDOWNDELAY
STOPOPTS="-a -A -s"
stop)
if [ -n "$SHUTDOWNDELAY" ]; then
SHUTDOWNDEL
fi
"$bindir"
;;