Veeam Backup Jobs

[Khue]

Hey everyone,

Recently acquired Veeam to replace an existing backup solution. I am working on figuring out the best way to construct backup jobs. Space and retention are my biggest concerns at the moment as my repository space comes at a premium. I have enough for my environment, but I have to offload to tape for long term retention. I want to make sure I am getting the best compression possible and after doing some research, grouping like VMs seems to be the best way to go (correct me if I am wrong).

I have about 62 total VMs. 43 of the VMs are Win2k8, 15 are Win2k3, and the rest are Ubuntu LTS of various versions. All servers are based off a template. What attributes should I look at to group VMs into backup jobs? I am pretty sure I don't want to just throw all 62 VMs into the same backup job.

[Gostev]

Hi, I would actually do exactly that (throw all VMs into the single job), unless you want different RPOs for some VMs, or if the resulting backup file will be too big to fit the backup repository (if you have a few small ones). Other than that, there is really no reason not to put all VMs into the same job for simplified management and better deduplication ratio. Thanks!

[Khue]

Even with varying server sizes? I would say 90% of my servers are around 60-100 gigs but I have a few outliers that are upwards of 1.5 TBs (approximately 3).

[Gostev]

That's right. Server size makes absolutely no difference. There were some considerations in earlier versions, but I will not confuse you with those, because they became irrelevant after we have introduced parallel processing. Thanks!

[Khue]

So I put together a basic list and apparently I have some storage constraints. Our old backup system was going directly to tape which was fine because we had vast amounts of LTO5 media. I have approximately 5Tbs of backups to get through according to the job estimation. Now obviously there's going to be some dedupe and compression of white space but without running the job first, it will be hard to tell. Does dedupe happen across the repository or the individual job?

[Gostev]

Individual jobs.