Scheduler Feature Request

[sandrola]

Hello
I have two Feature Request regarding scheduler:
The first one is for this situation:
We have actually 120 Jobs running on a Veeam Backup & Replication Server(VM) with 6 physical proxy server. If we set the jobs to start let's say at 20:00, the CPU of the VM jump to 100% (reclamed was 260%) and it fails to complete some Guest-OS related functionality (e.g. SQL Log truncation).
At the moment I have a Poweshell Script that "rebalance" the Jobs scheduled run time in 10min steps (from 20:00 - 4:00). So the CPU load is stable over the time.
The cons of this solution is that the script must be executed after every new Job is Applied.

It would be great to have as alterative to the static scheduler, Backup windows category with allowed backup times. Exemple:
Category 1 -> 20:00 to 0:00
Category 2 -> 0:00 to 4:00
Veeam should then start the Jobs in Step automatically during this time for the given category.

The second one is a Maintenance Windows that can be set globally over all Jobs. This would be usefull is there are maintenance on the infrastructure and no Jobs should be done during this time. As example Windows Patching once a month. Actually we have to activate the Job maintenance window via powershell script before the Patch Window and deactivate after the Patch windows. As alternative a Year-calendar (actually week) for the maintenance windows could also help.

many Thanks
Sandro

[foggy]

Hi Sandro, thank you for the feedback, but could you please clarify a couple of things? Do you mean CPU load on the Veeam B&R server itself? Is it also used as a proxy server (the default proxy is automatically assigned to the backup server)? Does it meet the system requirements in terms of CPU and memory? Basically, if you control the max number of concurrent tasks for proxy servers and system requirements are met, you can start all the jobs simultaneously and automatic scheduling will do its job.

As for the second request, then here's an existing thread about this.

[sandrola]

Hello
every proxy server (physical) has 36 cores and 128 GB RAM. We have 6 Proxy in 2 datacenter. The Veeam B&R VM has 16 vCPU and 24 GB RAM. This would mean 30-40 concurrent Job. The Repositories are configured for 36 parallel task each. We have 2 repositories for each Proxy. This would mean 72 task / proxy, 216 Task per datacenter. Is this limit to high? The problem at the Veeam B&R Server was the CPU and not the memory.
The default Backup Proxy is also present on the VM. The proxy selection in the Job is automatic. Could this be the problem? Can this proxy be disabled?

[foggy]

You're good in terms of RAM, but 1 proxy core is required for each concurrent task, so I'd limit the number of concurrent tasks per proxy at least to 34. As for the default proxy, yes, you can disable it in case you want to avoid it being selected by the jobs.

[Gostev]

It is a valid issue, but I think we will solve it differently. It should be trivial for us not to assign more tasks to the proxy if its CPU load is already sitting at 100%. This is something we planned for quite some time now, so thanks for sharing a validation of the need. Meanwhile, as foggy already recommended, please just reduce the amount of concurrent task slots in your backup proxy settings.

I will discuss adding the ability to disable all jobs at certain times with the UI folks. Should this include restores too?

[sandrola]

Hi Gostev, many Thanks for the answer. Do you refer to the maintenance window? Then yes, restores should be excluded as well.