Status: New
Owner: ----
New issue 788 by [email protected]: Wrong node execution of
post-migrate hook on non-DRBD disktype VMs
http://code.google.com/p/ganeti/issues/detail?id=788
There is a problem with the hook execution logic on instances with
sharedfile disktype. Due to it's nature, sharedfile (blockdev too) vms have
no secondary nodes so a migration/failover hook will be executed on the
wrong node. Because the INSTANCE_PRIMARY and INSTANCE_SECONDARY variables
are populated before the migration, the hook is only executed on the old
primary node and not on the one the instance it will be running after the
operation.
We have tested this on Debian Wheezy and Squeeze with Ganeti 2.9.3 and
2.4.2 respectively, but i'm guessing this affect every version.
We can see the hook execution log on the node-daemon.log on the old primary
node, but not on the new one.
The hook should be executed on the node that is primary after the migration
has taken place. Ideally, it should be executed on the old primary too in
case there needs to be a cleanup on the old node.
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
