dbsync error with unique constraint violation.
I have a botched update where we have a host that is getting the following error:
xapi: [debug||0 ||dbsync] dbsync caught an exception: INTERNAL_ERROR: [ Db_exn.Uniqueness_constraint_violation("VM", "uuid", "c8e359da-2964-451d-af40-802cabbdf889") ]
- yum update shows everything is up-to-date.
- There are no local SRs.
- Attempting to eject the host from the pool fails.
- The network interface configurations are whacked. Reset, and can't get bonds back up correctly either.
The host is configured to run DRBD for a Lizard-HA configured iSCSI SAN setup. (Which is not starting up).
It's like you have 2 VMs with the same UUID, which shouldn't happen in the first place
uuid ( RO) : c8e359da-2964-451d-af40-802cabbdf889 name-label ( RW): Control domain on host: xen1 power-state ( RO): halted
xe vm-list is-control-domain=true uuid ( RO) : 80506235-8ae0-4075-88d3-3c2f605d9b58 name-label ( RW): Control domain on host: xen2 power-state ( RO): running uuid ( RO) : c8e359da-2964-451d-af40-802cabbdf889 name-label ( RW): Control domain on host: xen1 power-state ( RO): halted uuid ( RO) : 059124f9-a845-498d-bbea-db8ac1b5ac70 name-label ( RW): Control domain on host: xen3 power-state ( RO): running
Okay so you have a host that you should consider dead to remove it from the pool (xen1), reinstall and re-join:
xe host-declare-dead uuid=<host_uuid>
xe host-forget uuid=<host_uuid> --force
Then, you can reinstall it and re-join the pool.
What's the master right now?
The master is system xen2, the broken system is xen1
Can this be done without erasing the iscsi/drbd configs?
I have no idea, I never used HA Lizard myself, but at least I can tell this should solve your current XAPI DB issue.
It's worth the attempt.
Maybe I can get it running standalone and work from there. I'm accessing everything remotely, no people there in the office.