Best Practice Question: Replicate Disk based on day of week

[micoolpaul]

Morning all,

Appreciate it's a strange request but here's the scenario. I've got a server that intermittently fails to use CBT (it's an old Windows Server 2003 R2 that used to be physical, don't judge please I'm retiring it ASAP). It is a SQL Server that has over 1TB of documents stored as files in standard Windows folders.

Now whenever CBT fails, my backup takes all night just for the one server, since it's not the only server in the job the replicas run into the day and it's painful at times. The attachments come in via email and we have our own third party cloud mail storage so in the event of a disaster our email records would remain intact so the documents don't bother me if I lost a few days worth.

Has anyone managed to/is it supported to do the following:

Replicate C drive and any DB disks & appropriate logs on a nightly backup.
Replicate Data drive Friday night to update the week. It's normally quick but if CBT fails it's got the weekend to process the data and if CBT fails during the week it's about 50-60GB total to process instead of 1-1.2TB of data.

If anyone else was facing this problem how did you set about eliminating it?

For extra info I'm using 100Mbit MPLS with WAN accelerator already so I can't throw much more at it that way.

Thanks,
Michael

[widmerkarl]

Hi Michael

Did you try that with Disk Exlusion?

https://helpcenter.veeam.com/docs/backu ... tml?ver=95

On VM leven you can exlcude disks you don't want to backup (or better to include disks you wan't to backup).

Probably you can edit the daily job to only include the system and DB disks. Then you create a new job for weekly execution for this server and only include the data disk.

[micoolpaul]

Hi,

Yeah I've used that in the past but the in terms of stitching together has anyone tried and had any success in having two replica jobs tied to the same replica VM, one just updating in my instance a data disk on a weekly basis and then the DB and core system disks updating on a daily basis?

Wasn't sure how Veeam would handle something like this when it came to recovery or failover. Otherwise it'd be manually orchestrated failover wouldn't it tying together the disks to get a boot?

Thanks,
Michael

[foggy]

having two replica jobs tied to the same replica VM

I don't think this is a good idea.

[micoolpaul]

Hi Foggy,

That's what I was expecting but I'm not sure how best to proceed with this.

Any suggestions on this would be greatly appreciated. It's normally fine but the moment it needs to recreate fingerprints or CBT data is invalid the replication takes about 2-3 days.

Sorry if it sounds like I'm pinning this on Veeam as a problem, that's not my intention. Just struggling to find what the best case would be for this.

Thanks,
Michael

[foggy]

Using two jobs with excluded disks and different schedules seems to be the only workaround.

[micoolpaul]

Okay and then if it came to doing a complete failover I'd just need to manually attach the Vdisks together?

Thanks!

[foggy]

I'm just not sure about consistency between them, since they're going to be from different time.

[micoolpaul]

Well the mindset is to do All the drives apart from the raw file attachments drive in the one core job and then just a weekly file attachments. If you don't think it's something Veeam can do fair enough I'll live with the longer rebuilds when CBT fails etc but just hoping to save time where possible.

[foggy]

It doesn't come down to Veeam B&R specifically, but to the whole idea of such setup. You will have both backups consistent per se, but spread out in time, which might cause issues.

[micoolpaul]

Not a problem, completely understand, just didn't want to be missing a trick. Thank you for your input