Confirming some definitions: jobs/tasks etc

[cosmik]

From the requirements at https://helpcenter.veeam.com/docs/backu ... ml?ver=110 I'd like to confirm that I really understand what it is about:
* for backup server "4 GB RAM plus 500 MB RAM for each enabled job". Obviously job is whatever is under "Jobs" in the UI. Therefore the first job approximately needs an additional 500Mb and each one another 500M right?
* for the backup proxy there is the notion of "concurrent task". Does this correspond to the number of disks the proxy is processing? Are there other meanings to this as well?
* for the backup repo there is a mention of memory per "... concurrently processed machine or file share". I have a single job containing say 10 VMs. What is the meaning of concurrently processed machine here?

Thanks in advance!

[HannesK]

Hello,

backup server: yes, correct
proxy: there is one task per disk for VMware / Hyper-V, yes.
repository: machine = VM in that case. the number of machines depends on how many machines your proxies concurrently process (depends on number of disks).

What number of VMs are we talking about here? There is a sizing tool that might help if we talk about hundreds / thousands of VMs. For smaller installations, one can use easier rules of thumb

Best regards,
Hannes

[cosmik]

What number of VMs are we talking about here? There is a sizing tool that might help if we talk about hundreds / thousands of VMs. For smaller installations, one can use easier rules of thumb

Very small setup here: 15-20 VMs. A mix of Windows 10, 7, XP (yes even XP), Server 2019, CentOS 7/8, Debian 32 and 64-bit... This is a heterogeneous mix as far as it goes (which will present an issue with my copy jobs later on). From some test runs I've seen that 3-4 VMs might be accessible at a time at various moments of backup...

[HannesK]

I cannot imagine that anyone can get a smaller server than 32GB RAM and less than 8 cores today. That's enough for 15-20 VMs (except they are massively larger than average)

in case you thought about putting the software in a VM: it's recommended to have a backup environment separated from production. that prevents from chicken / egg issues

[cosmik]

I cannot imagine that anyone can get a smaller server than 32GB RAM and less than 8 cores today. That's enough for 15-20 VMs (except they are massively larger than average)

Cluster is 2x 16 core rigs, 64Gb each (over dimensioned, but I'm not gonna buy anything for the next decade+).

in case you thought about putting the software in a VM: it's recommended to have a backup environment separated from production. that prevents from chicken / egg issues

I do have the software as a VM, with the notable exception of the repo(s), which will end up a hardened physical installs with immutable backups.

[HannesK]

ah well, I remember your other post. I would go then with one backup server VM with 6 cores and 12 GB RAM.

[cosmik]

You're da man! Thanks!