Interesting Job Scheduling Problem

[tthomas1@ebsco.com]

Hello,

I have an interesting job scheduling problem and would love feedback on potential solutions.

Scenario: I'm running a full backup on a physical server every Friday night that runs just shy of 72 hours. It runs an integrity check afterwards which causes it to run >3 days, blocking the Monday incremental job. I really need the Monday incremental to run.

First: I can change the integrity check to run with the Monday incremental. That way the full backup will still stop just shy of 72 hours and the Monday incremental will run. My question is this: does the integrity check ONLY check the incremental, or does it also go back and check the integrity of the full backup the incremental is based on? If not, it would kind of defeat the purpose of the integrity check. If this is NOT a viable solution to my issue, does anyone have a suggestion?

Second: even though the full backup (minus integrity check) finishes in just shy of 72 hours, that will likely change very soon and spill over past 72 hours. My solution would be to reschedule JUST MONDAY's incremental backup an hour later. However, I don't see a way to do that. Can two separate jobs be run that affect the same backup chain? I could have a single job for Monday incremental scheduled an hour later, and the main job for the other nights of the week. Would love any feedback on whether that is possible!

Thank you in advance for any help offered!

[PTide]

Hi,

Which agent is that and which job mode are you using?

Thanks!

[tthomas1@ebsco.com]

Hi PTide,

This is the Windows agent and it is a standard backup job. Full on Friday and incrementals every other night of the week (although as I mentioned, because full runs nearly 3 days, there is no incremental on Saturday or Sunday).

[PTide]

My question is this: does the integrity check ONLY check the incremental, or does it also go back and check the integrity of the full backup the incremental is based on?

It verifies the latest restore point in the backup chain (restore point created with the current backup job session — the session during which the health check is performed. "Restore point" stands for "the latest state", not for "the latest file".

My solution would be to reschedule JUST MONDAY's incremental backup an hour later. However, I don't see a way to do that.

You can use a script to trigger backup manually on Mondays (just disable Monday start).

Another question - do you run synthetic full or active full? If the latter, then I' suggest to try synthetic and see if it will go faster.

P.S. Do you use VBR to manage the agent or it is just a standalone installation?

Thanks!

[tthomas1@ebsco.com]

Thanks PTide! We run synthetic full already, and we use VBR to manage the jobs.

The script idea looks perfect! Main question: will it work for a job managed from a VBR server?

[tthomas1@ebsco.com]

All,

I'm starting to realize PowerShell is my friend here. Apparently there is a relatively new command called Start-VBRBackupHealthCheck. It's not well documented, but I assume it can be used to manually run the health check on the full backup. I can also create a script with Start-VBRJob to manually start the Monday incremental.

Main question if anyone knows: if indeed the health check can be run separately, can I schedule it to run Monday night after the Friday full finishes, and then have the Monday night incremental job run at basically the same time without them interfering with each other? Or will they run over each other and cause issues?

[PTide]

No, unfortunately. You cannot control an agent if it's managed by VBR, and you cannot trigger it (or healthcheck) with PS. Veeam Agent for Microsoft Windows cmdlets will not work for jobs that were created using Veeam Agent management functionality. The ability to manage Veeam Agent management jobs will be provided in future versions of Veeam PowerShell.

Another workaround that came to my mind was to schedule the job to run periodically (evey X hours) and specify a backup window in such way that the job will start one hour later on Monday.

Thanks!

[tthomas1@ebsco.com]

Thanks PTide! I am actually testing a VBR PowerShell script now that is creating an incremental. If it works, I believe my solution will be to run the full backup + health check as normal, which is approximately 72 hours + 6 hours, and then schedule the incremental separately via PowerShell to run early Tuesday morning to cover Monday's data.