Feature Request: allow excluded (OU) protection group agents to be part of another backup job

[edison5000]

case# 05604784

or perhaps, "allow 'excluded' agents in protection group to get backed up as part of another protection group".

We have agents in a protection group that are part of an OU, which we are not allowed to move from that OU - and they need different backup job type/parameters than the other machines in the same OU. Arturo in support was able to duplicate.

This means every system in that protection group OU will either
a) get whatever backup job affects the whole protection group, or
b) will potentially get left out of the job, if you exclude agents from a protection group by using the 'exclude' function in the inventory dialog.

But - although excluding the agent will strangely permit you to add that agent to another protection group at the same time, backup jobs that run in workstation mode - > managed by agent do not allow more than 1 job per agent. So once the agent is in a 2nd protection group, even if it has been excluded from the first, a different backup job will fail with 'sorry, this agent already belongs to another job'.

A weird workaround we and Arturo discovered, is, you can try to trick Veeam into selecting the agents you want backed up differently, if you get them recognized first by the protection group/backup job you want them in, before the protection group with the OU in it sees them. This will then do the backup job we want on the select agents, and those agents in the OU protection group will get the 'sorry, these machines already belong to a backup job' warning when that job runs. But this method is not a good long term solution, is unreliable, and only seems to work if we start out knowing this from the get go (as opposed to machines that already belong to an OU in a protection group).

Sorry if I was not clear, but it is easily duplicated. : )

Thanks!

[Mildur]

Hi Edison

Thanks for the case number. I looked at the internal case comments and you are right, that shouldn't work in that way.
Sounds like a bug to me. It must be possible to add an excluded workstation agent from the OU protection group to another and then do backups with another job. Please keep working with the support engineer. I'll watch the case from our side (PM) and forward it if necessary. It seems that your support engineer will probably also open a feature request to solve this behaviour.

Thank you
Fabian

[edison5000]

Thx Fabian. No sir. You cannot get another job in workstation/managed by agent mode, to backup an excluded agent (hence the feature request : ) So maybe bug.

To reiterate -

The only way to insure an agent that belongs to two protection groups, but is part of an OU in one of them, because you want to apply a different job to only that 1 agent, currently - is to

1. make sure pc of choice does not exist anywhere, historically, in Veeam
2. create protection group A. Add A.D. machine 1 individually to protection group A. Create and apply backup policy A.
3. create protection group B. Add the OU that machine 1 is part of to protection group B. Create and apply backup policy B.

Now, the policy A that is applied to machine 1 in prot group A will work, and, be different from the job applied to prot group B. When policy B arrives at machine 1 in job B, it will skip over it with the warning "you cannot have an agent in two jobs", or similar. This is likely working as designed.

But - if you do the above in reverse order, such that the OU with machine 1 in it exists first, it won't work, because it will already belong to another job. And indeed, the same thing happens even if you 'exclude' machine 1 from prot group B.

fyi we are on B&R 11a and all agents are 5.0.

No other info so far. Will post again if I hear something new.

[Mildur]

Hi Edison

Yes, it's definitely unexpected. RnD is looking into the case.
The latest reply from support with resetting the configuration database looks promising as a workaround.
Have you already tried that?

Thanks
Fabian

[edison5000]

Hi again Fabian. The machine in question now starts a new backup with no problem, but randomly fails in the middle. This is after resetting the database with that reg key, as well as removing it from the B&R configuration and readding it. Per the agent logs, Arturo says he detected a bad block on the agent pc, so I'm trying to repair that first. The other potential reason it doesn't seem to be completing is that it says repo is not ready, but I find that really odd because - barring a complete lack of server resources or storage, if there is space available in the target repo, why would Veeam say the repo has no resources available after the backup has begun? At that point, 'backup infrastructure resources' have *already been assigned* to that agent - it should no longer be waiting for any. Will report back after more testing.

[edison5000]

hello again. This is still a problem in v12, with v6 agents. It's not urgent, but thought I would let you know.