Maximum concurrent job

[eantonio]

Case: 00927631

I use HP DL380p Gen 8 server as my Veeam backup server and also my initial repository. It has 8 cores and 64 GB RAM.

I have multiple Veeam backup jobs setup and my backup repository is the bottleneck. I have configured it to do a max of 6 jobs. What do you recommend as my maximum concurrent job that I can run without affecting performance. Thank you.

[foggy]

Emil, in such setup this server holds both proxy and repository roles and should meet the combined system requirements, which can be used to calculate the max number of concurrently running jobs.

I have multiple Veeam backup jobs setup and my backup repository is the bottleneck.

Do you mean Target is identified as the bottleneck for your jobs?

[eantonio]

Yes, the veeam backups are in pending status saying the waiting for Backup Repository.

So if I have 64 GB RAM, can I do a maximum of 16 concurrent jobs to my backup repository?

[eantonio]

Is the limitation more on the software and amount of RAM available? and not the disk hardware itself?

[foggy]

Yes, the veeam backups are in pending status saying the waiting for Backup Repository.

I meant bottleneck stats reported by the job session (if you right-click the job and select Statistics).

So if I have 64 GB RAM, can I do a maximum of 16 concurrent jobs to my backup repository?

While you may be ok in terms of RAM, note that each task requires a single CPU core and your disks should be capable of the corresponding amount of I/O.

[eantonio]

Yes, the stats show the bottleneck is the Target.

[foggy]

Then increasing the number of concurrently running jobs will affect its performance even more.

[eantonio]

What do you suggest I do to resolve this issue?

[Andreas Neufert]

What issue?

If your target storage isn't fast enough to manage the load you can replace it with a faster storage.

Ok let's have a look at the parameters.

8 Core can officially run 4 proxy and 4 repo task slots.
Start with this configuration.
Bottleneck at Active Full (first job run) doesnt matter. Check the incremental runs.
If it us still the target and you are not OK with overall prozessing time, you can reduce repository tasks to 2.
As well you can change job from Reverse Incremental to a Forward Incremental chain to reduce load on target storage.

In Target Storage you can change Raid 6 to Raid 5 to increase performance.

Therewill be in any way a bottle neck. The question is if you are OK with the prozessing time or not. Then work on the shown bottleneck if you want to decrease prozessing time.