Feature request: repository drain-stop

[ncolt]

Hello,
I'd like to be able to put my Windows repositories into a sort of maintenance mode so they can be patched. This wasn't a problem when I just ran backup jobs overnight but now i run backup copy jobs during the day and am also starting replication so will be looking at having to cancel/disable jobs so that I can reboot the server as there will be few times during the day when the repositories are doing nothing.

I would love to see a drain-stop feature which let existing jobs run their course and not start new jobs.

Neil

[PTide]

Hi,

<...>which let existing jobs run their course and not start new jobs.

Do you mean that the "drain-stop" feature should let you pause the running job in such manner that it should be able to resume the backup process from the point where it has been paused? Please elaborate.

Thank you.

[ncolt]

Hi, no my initial request is that they could run their course.

To be able to pause jobs could be useful as well but in this case a job could start up just as you are about to reboot the server if you hadn't paused it.

[PTide]

is that they could run their course.

Yes, I got that, but I still don't quite understand what do you mean by "run their course" in the context of "reboot" and "maintenance mode", sorry.
Could you describe your use case in steps please?

Thank you.

[ncolt]

The monthly patch is due
You put the repository into maintenance mode
Existing jobs carry on running
No new scheduled jobs start running until you remove it from maintenance mode
Existing jobs finish - maybe you get an email?
Patch the server, reboot
Remove repository from maintenance mode - possibly takes itself out of maintenance mode due to reboot

Any jobs which would have run need to be paused otherwise they may miss the schedule

[PTide]

Ok, sounds reasonable, thanks for the suggestion.

[dellock6]

The described feature is exactly what Scale-Out Backup Repository and its Maintenance Mode will allow to do in v9: you can create a scale-out with just one repository, put it in maintenance mode, and because it's the only available extent in the scale-out there will be no failover to another repository. Running taks will complete before entering the maintenance mode, and after the operations you will be able to disable maintenance mode. Only thing missing from your requests is the auto-exit from maintenance, and it makes sense since our software cannot know the reason for the maintenance, so we cannot decide when to get out of it.

[ncolt]

Excellent, thanks Luca.