Urgent: how to stop a backup task
-
One of our servers have failed, and it is busy exporting. I need to stop that task as the server is in production and it's taking too much of a beating. Any advice? I cannot seem to click on cancel of the task under tasks screen
-
You need to restart toolstack to cancel an export.
-
@olivierlambert will it also cancel the running backup job or just cancel the current backup? It appears our incremental backups started doing full backups today, and this has now caused a bottleneck on one of our primary servers as it was too busy with a file level backup and I think malware scanning so the IO timed out. Issue now is that I can boot the server but because all of the services are starting up (we have about 250 hosting accounts on it) the same occurs. I want to stop the running tasks but also the backup job so that I can focus on the 1 thing.
I assume restart toolstack on both sending and receoving server (it's a CR) - Should I restart XO as well?
-
If you restart the toolstack, the current export will be seen as "cancelled" by Xen Orchestra, but it will continue then on the next VM in the backup job.
If you want to cancel both, disable the job in XO then restart
xo-server. Then, restart toolstack on the host. -
@olivierlambert thank you Olivier and thank you for your prompt response. In hindsight this was all due to a combination of bad planning. Our full backup interval is set to 20 days, and this then led to full backups being run this week and that along with a 2nd set of backups + malware scanning led to the IO being overloaded and stopping / restarting the service.
To avoid this scenario, and I will be default assume this cannot work out of the box, but is there a way to set a schedule around the full backup interval? We would like to set these backups to never run on weekdays. If not, what can you recommend we do to retain a delta chain but also do full backups over a weekend where the schedule is set to run daily?
Once again thank you for everything you are doing for the community!
-
@mauzilla It's been a few years, but this seems like a common issue. How have you worked around it?
-
@andrewperry I think healthchecks is the answer here. We're not backing up vms but rather incremental replication with health checks. If a vm does not fail health checks I cannot see a reason for a full backup unless it becomes a snapshot chain issue. Olivier might be able to provide better insights here, we're in process of implementing the above will keep you posted
-
I would go for healthcheck and/or chained backups that are now available
-
@olivierlambert do you have a link where we can read up on chained backups?
-
The latest release blog post, in the Backup section: https://xen-orchestra.com/blog/xen-orchestra-5-99/
