Waiting for backup infrastructure resources avaiability...

[mkretzer]

Hello,

since we use per-VM backup chains for our backup copy target we get "Waiting for backup infrastructure resources avaiability" more often in the copy job.

I do not understand why. For example right now the following is running:

Primary (source) Repo (12 concurrent tasks set, per-VM backup chains off)_
- 1x Synthetic Full task

Target Repo 1 (8 concurrent tasks set, per-VM backup chains off, on the same server as Repo 2):
- 1x Health Check

Target Repo 2 (6 concurrent tasks set, per-VM backup chains on, on the same server as Repo 1):
- 1x Backup Copy

Now the backup copy on Repo 2 is halting after each VM with "Waiting for backup infrastructure resources avaiability...". How can we find out what ressources are the issue?

Markus

[foggy]

Markus, this is not due to per-VM backup chains setting itself but due to parallel processing available now in the backup copy job. You do not have enough free slots on the Target Repo 2 to process all VMs the source repository is ready to supply: 6 on the target versus 11 on the source (1 slot is taken by the synthetic full).

[mkretzer]

Jim, i would expected that behavoure but what we see is that no data is transferred between the VM copys, it just pauses with the message. Is there any way to find out the current usage (how many processes are counted) of all the components so we can find out the "bottleneck"?

We are quite sure it is not a classic IO, CPU or Network bottleneck as all system statistics look good. We think something is counted wrong by Veeam or some counter did not get reset so it "thinks" it has no ressources.

[foggy]

Feel free to ask our support engineers for assistance with that.

[mkretzer]

Will do the next time it happens! Thank you!

[mkretzer]

Jim, i have read some other threads and i think i know whats "wrong". Since we only have per-VM backup files on the target there is no paralell processing on the copy job correct? And then the "waiting for backup infrastructure..." is coming between every VM but only for a short while (which might be normal)...

So if we enable per-VM backup files on the source this should change if my theory is correct?

[foggy]

No, per-VM chains do not have any effect here. You just need a direct mode (not WAN accelerated) job and enough task slots available.