yahoo-eng-team team mailing list archive

[Sylvain Bauza <sbauza@xxxxxxx>]

Not sure it's a Nova problem but rather a MySQLdb problem as per the
log. Could you please try to see the config differences between the 13
good nodes and this one ?


** Changed in: nova
       Status: New => Opinion

** Changed in: nova
       Status: Opinion => Incomplete

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1454410

Title:
  Juno HA Deployment, percona on VIP, killed leader, compute did not
  reconnect

Status in OpenStack Compute (Nova):
  Incomplete

Bug description:
  ii  nova-common                            1:2014.2.2-0ubuntu1~cloud0            all          OpenStack Compute - common files
  ii  nova-compute                           1:2014.2.2-0ubuntu1~cloud0            all          OpenStack Compute - compute node base
  ii  nova-compute-kvm                       1:2014.2.2-0ubuntu1~cloud0            all          OpenStack Compute - compute node (KVM)

  I have a 14 compute node deployment, my VIP percona leader was axed
  and one of the compute services did not reconnect.  13 did, so clearly
  there is retry logic already involved that is working in most cases.

  Will attach full stack trace, The missing log entry that all the good
  units have seems to be this:

  2015-05-12 19:55:52.867 229965 ERROR nova.servicegroup.drivers.db [-]
  Recovered model server connection!

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1454410/+subscriptions