Question about conflicts with 2 jobs backing up the same VM?

[dpcallais]

Hi,

I want to create two backup jobs that will both backup the same VM with the following setup:

Backup Job 1: Forever incremental backups taken once per day

Backup Job 2: A full backup is taken once per day and is then followed by incremental backups taken every hour for the rest of the day

Would there be any conflicts with running these two backup jobs against the same VM? My primary concern is that I want to ensure that when Backup Job 1 incremental runs, it is backing up all of the changes since the last time Backup Job 1 ran on the previous day instead of just the changes that have taken place since Backup Job 2 ran an hour ago. Also, would there be any conflicts if Backup Job 1 & 2 are both running at the same time?

Thanks,

[PTide]

Hi,

Would there be any conflicts with running these two backup jobs against the same VM?

There will be no conflicts until jobs run during different time periods.

My primary concern is that I want to ensure that when Backup Job 1 incremental runs, it is backing up all of the changes since the last time Backup Job 1 ran on the previous day instead of just the changes that have taken place since Backup Job 2 ran an hour ago.

Backup chains are independent so Job1 will grab all the changes that have happened since last Job1 session.

Also, would there be any conflicts if Backup Job 1 & 2 are both running at the same time?

Each Job takes a snapshot of your VM, so running two Jobs simultaneously means that there will be two snapshots of the same VM on your production storage which will result in a performance decrease - your VM will slow down.

Have you considered using Backup copy job instead of using forever-forward job? It seems that this is exactly what you need.

Thank you.

[timmi2704]

Hi,
any information about CBT usage when 2 different jobs process the same VM independently?
How will this behave?
Thanks

[foggy]

There're no issues in having multiple jobs pointing to the same VM, CBT is not impacted either (this is covered in the sticky FAQ topic).

[iami]

I am working with an existing configuration where a single server is in multiple backup jobs. There is a monthly job that triggers on the first Sunday of each month with 12 restore points, a weekly job that starts Friday evening with 4 restore points and a daily that runs MTWT with 5 restore points. My understanding is that backup chains are per backup job and that CBT will only backup blocks that have changed since the previous backup. Will this work or will all the backup chains be missing data? For example, if data is changed on Friday then the weekly incremental will capture that but the next daily on Monday won't, hence the change will never make it into the daily backup chain. Or will the changes captured on Friday by the weekly job also be applied to the backup chain of the daily job?

[DGrinev]

Hi Ian and welcome to the community!

You can backup or replicate the same VMs in a different jobs as CBT has been designed to handle such usage scenarios. Thanks!

[iami]

Perfect, thanks. Conceptually I had envisioned CBT to be a bitmap with a single bit per block with the bit being zero for no change since previous backup and 1 when changed. That would have meant that it wouldn't have supported multiple backup jobs (or at least they would interfere with each other). So obviously it is more complicated than that and armed with that information and for those interested this excellent Veeam article explains it in more detail. To summarise, VMware keeps a changeID for each block rather than a single bit which allows software to ask for all blocks that have changed between two changeIDs. These can be thought of as timestamps with the internal implementation by VMware hidden. Very neat.

I do agree with PTide that the Backup Copy functionality seems more appropriate for creating a different backup chain for the same VM which is what we are doing by having multiple backup jobs. What is Veeam's recommendation?

[DGrinev]

The backup copy job was designed to create separate backup chains.
The highly recommended way of data protection described by the 3-2-1 Rule.
So one of the benefits is reducing workload on the production, since BCJ is using the data blocks from the backups retained in the repository. Thanks!

[rntguy]

Hello,

If there are two separate jobs backing up the same VM, where a Copy Job doesn't make sense, for example, how does Veeam handle CBT and incrementals?

For example

Day 1:
VMa is backed up by Job1 (Full)
VMa is backed up by Job2 (Full)

Day 2:
VMa is backed up by Job1 (incremental) - it sees CBT data and goes very quickly to do an incremental
VMa is backed up by Job2 (incremental) - is CBT data still available or does it rescan to find changes manually and slowly?

Does any of this matter if VMa has SQL or exchange databases? I know transaction logs might have been truncated by one of the jobs if set up that way.

Thoughts? For now suffice it to say there's a good enough reason to have 2 jobs. Need to know how the changes are handled.

Thanks

[DGrinev]

Hi,

No worries as CBT was designed to handle such usage scenarios.
Please review this thread for additional information. Thanks!

[rntguy]

thank. that answered my question.