Required backup infrastructure resources have been assigned

[kirkgj]

Its seems that Endpoint Backup goes through two separate rounds of assigning infrastructure resources for each backup. The first round is to complete the backup, the second round is to merge incremental forever.

If I use a Backup and Replication server repository with a Load limit of say 10 concurrent tasks and I have more than 10 Endpoint backups running at a time, the endpoint backup that finishes the backup first then has to wait for in the rest of the queue to complete before the first one can then move on to the merge process. This results in all backups having to wait in the middle of their backup process for resources to be assigned again. The more backups are in the queue, the longer all Endpoint backups get stuck in the middle of their backups waiting.

It would be nice if the assigned resources could stay with the running backups for the entire backup process (backup and merge). Or better yet, complete the backup at the client and let the Backup and Replication server take care of the merge without the Endpoint client being involved.

Kirk

[Dima P.]

Hello Kirk,

That’s interesting thanks for sharing! I’ll ask our folks to check this behavior.

[kirkgj]

I was able to capture this scenario as it happened.
In this situation the Veeam Endpoint repository has been setup with a resource limit of 9 concurrent jobs. In the first image we can see that the backup and replication server shows EEI-NB215 is 10th in the order of jobs but it actually obtained resources as shown in the Endpoint Backup image and was able to get some of the way through a backup before having to wait for resources again.



Now, in the second image, we can see the backup and replication server shows EEI-NB215 is now 9th in order of jobs but it is still waiting halfway through the backup for the second resource assignment as shown in the Endpoint backup image. To me this means that another job after EEI-NB215 is now using available resources, since EEI-NB215 is now in the top 9 jobs and not able to proceed yet.



There must be something going on with the resource queue and how it assigns resources to the Endpoint Backups. It would be nice to have the current backup hold its resources for the entire backup process and not have to go back to the queue to get reassigned.

Kirk

[Dima P.]

Hi Kirk,

Thanks for the screenshots. I had a conversation regarding this behavior today. It's designed this way to provide the ability to create a backup on a first place for multiple backups jobs running at a same time, leaving the rest of the time to merge process. So, once backup is completed – merge takes place, however, if another job must create a backup, it locks the available slot immediately. As a result, you get a backup file created right on time and then merge kicks in.

[kirkgj]

Thanks for the response. If my feedback and example could result in a new feature to hold the queue assignment or a better queue management in the future, that would be most helpful. It seems like the end user is more interested in a quicker end to end backup experience as opposed to a better chance of running the backup on time.

Kirk

[Dima P.]

Kirk,

I agree that this might be a bit confusing from an end-user perspective. I’ll definitely discuss it with the team – thanks! Out of curiosity, are you backing up end-user’s desktops/laptops or the screenshot lists some of your ‘last remaining’ physical servers?

[kirkgj]

Dima,
I appreciate you following up with this. The screenshots are a list of our end user laptop backups that were running. In this case, my laptop is EEI-NB215 and those are the Endpoint Backup sub-screenshots from my laptop. Veeam Endpoint is working out great in our environment. I just want to help make it even better.

Kirk

[Dima P.]

Thanks for the reply Kirk!