Our Ignition server is installed on a VM with Windows 2016 server, 8Gb RAM, the ignition plateform is assigned 2Gb. We do a backup of the VM every day and we have a warning ClockDriftDetector because of it. It’s just one warning, but the deviation varies between 2s and 15s.
What is the impact on my applications?
Should I increase the memory of the ignition gateway even thought we never go over 50% of use?
Thank you
I would check into what type of memory is assigned via the VM. We've seen issues in the past if the memory isn't dedicated to the VM where it may not have enough memory at the instance that we needed it if it was a shared allocation in the system.
It is shared memory, but the server is oversized. Nothing else is using the resource, would that really have an impact?
I don’t think this is a memory issue. It seems more like an IO issue to me.
Backups typically cause a lot of IO, and Ignition writes quite some stuff to its internal DB (even if you don’t use historian or other SQL things).
If there’s a possiblity, you could try to lower the IO priority of the backup process. But that’s not always possible.
Do the operators notice it happening? Is there a possibility to execute it during a pause?
If the operators don’t complain about it, and you know the cause, I wouldn’t care too much about it.
That's interesting answer. We aren't yet in production, so I don't have user feedback yet. I will see if we can modify the priority of the backup.
Thank you