On Sun, Apr 4, 2021 at 4:57 PM <phunyguy(a)neverserio.us> wrote:
Hello, exactly as the title suggests, I did something stupid, and I am not sure how to
fix this. I also cannot be certain that this was the exact cause of the problem, as I was
troubleshooting another issue with a different VM refusing to start.
What I did by accident, was selected the hosted-engine VM, rather than the VM I was
attempting to troubleshoot, and changed the assigned network of the VM adapter to
something else, and of course, lost connectivity to the engine VM. This caused HA to kick
in and restart the engine, however, like the other VM, it is now refusing to start.
Can anyone give me any pointers? agent/broker/vdsm.log are not showing anything useful
that I can see. Thanks in advance.
You can try connecting to the engine VM's console with 'hosted-engine
--console' to see its current status.
If indeed its network configuration is broken, you can try something like:
1. Move to global maintenance
2. Keep a copy of its current configuration - should be in
3. Shutdown the engine vm
4. Edit your conf copy and fix the network issue. If you have older
you might want to check the engine db dump inside them - I see that we
have a table
vm_ovf_generations that might include this information (I do not know
it, just searched the
dump for 'HostedEngine' and 'xml').
5. Start the VM with 'hosted-engine --vm-start --vm-conf=file'
6. If all looks ok, edit the engine vm from the web admin ui and fix
the bug there as well
7. Make sure your saved conf is propagated to the shared storage and
local copies in /var
8. Stop global maintenance
9. Test - shutdown engine vm, try to start it
I think you should be able to find details for specific steps in
existing docs, this list's
archive, elsewhere on the net, etc.