Software keeps switching to wrong mode?

[danswartz]

I have a virtual CentOS 7 VM that acts as a backup repository for my normal B&R install. Obviously it can't back up to the repository itself, so I installed veeam agent for Linux and have it back up to a directly attached SSD. The problem: I installed it with no license, in workstation mode, and a test manual backup works just fine, as does the 6:00AM scheduled backup the next day. Within a day or so (not sure when), backups stop working, and when I enter 'veeamconfig ui', a number of command options are no long visible. Entering 'veeamconfig license show' yields this:

[root@nas ~]# veeamconfig license show
Error: This operation is not permitted in Managed by backup server mode.

The only way I can seem to 'fix' this is to uninstall, install and reconfigure. What the heck is going on???

[danswartz]

So looking at /var/log/veeam/veeamsvc.log, I found a smoking gun:

Code: Select all

[22.05.2019 21:00:26] <140199632484096> lpbcore|   Acquiring agent. VBR server: [pdc.druber.com]. Installation ID: [{f70c4fea-eb21-\
4728-b2d8-2af7240d0984}]. Reset: [0].
[22.05.2019 21:00:26] <140199632484096> lpbcore|     The agent is being switched to managed by agent mode.
[22.05.2019 21:00:26] <140199632484096> lpbcore|     Switching to standalone mode.
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Removing all backup policies.
[22.05.2019 21:00:26] <140199632484096> lpbcore| WARN|Backup job [BackupJob1] will be removed.
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Creating repository database accessor for local database.
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Update job [{4b926506-4ce4-4641-97c6-6b21224d68ca}] in crontab.
[22.05.2019 21:00:26] <140199632484096> lpbcore|         Job [{4b926506-4ce4-4641-97c6-6b21224d68ca}] is not found.
[22.05.2019 21:00:26] <140199632484096> lpbcore|         Removing job [{4b926506-4ce4-4641-97c6-6b21224d68ca}] from crontab.
[22.05.2019 21:00:26] <140199632484096>        |           Creating child process: crontab with arguments: -e
[22.05.2019 21:00:26] <140199632484096> lpbcore|           Job is presented in crontab and will deleted.
[22.05.2019 21:00:26] <140199632484096>        |           Creating child process: crontab with arguments: -
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Update job [{4b926506-4ce4-4641-97c6-6b21224d68ca}] in crontab. ok.
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Switching all VBR servers to a restore-only mode.
[22.05.2019 21:00:26] <140199632484096> lpbcore|       Switching all VBR servers to a restore-only mode. ok.
[22.05.2019 21:00:26] <140199632484096> lpbcore|     Switching to standalone mode. ok.

I saw serveral of the 'Acquiring agent' messages when everything was working, so I'm not sure that is an issue. That said, I'm puzzled as to why it is doing *anything* with pdc.druber.com (my B&R VM), since I could swear I did the install in standalone mode!

[danswartz]

I've deleted & reinstalled again. Deleted some random cruft that was left over. If this happens again, I will update this post. If someone at veeam understands what is biting me, I'd be most grateful

[PTide]

Hi,

It seems that VBR is trying to control the Agent instance that is installed on the repo server. That might occur if the agent is added to jobs or protection groups on VBR. Please check that and and remove the agent from those groups and jobs if needed.
Otherwise, kindly contact support and post your case ID.

Thanks!

[danswartz]

Ah, hmmm. I had a previous CentOS 7 backup repo VM that had gotten way too complicated. I backed it up to the regular repository, just so I could recover files from it if needed. It is currently sitting there in the Backups => Agents view. I'll purge that and see what happens...

[PTide]

You shouldn't purge backups, you should check if there are Agent jobs that contain the machine.

Thanks!

[danswartz]

Well, silly me. There was one of those too It's gone now. I'll post an update as things develop...

[danswartz]

Well, this is frustrating. I made sure there is NOTHING on the B&R server relating to this centos7 VM. Come downstairs this morning, and... 'Operation mode has changed'. veeamsvc.log says:

veeamsvc.log:[25.05.2019 21:00:10] <139785901324032> lpbcore| The agent is being switched to managed by agent mode.
veeamsvc.log:[25.05.2019 21:00:11] <139785909716736> lpbcore| The agent is being switched to managed by VBR mode. All current policies/jobs will be removed.

This is very odd, since I had had another centos7 (not VM but physical) that I was using the free linux SW for for months with no issues. The only thing I can think of is that this VM is in the B&R console as a backup repository, but then again, the physical centos7 host was also, and that worked just fine

EDIT: I did just now notice I had a leftover reference to 10.0.0.30 as an agent backup repository. I've now deleted that, but I'm not sure how that could have been germane, since as I said, I was running another centos7 physical host with 10.0.0.30 also acting as a linux backup repository. It's almost like being a VM is what is triggering this?

[PTide]

Ok, if the machine does not belong to any protection groups, jobs, or policies, then the only possible reason (besides it being a bug) is that the VM's IP previously belonged to some other machine which had been added to a PG/Policy/Job.

Let's see how it goes, and if the VM will keep going into managed mode I'd suggest to contact support team directly on that.

Thanks

[danswartz]

Aha, may have found it. nas.druber.com was still in the manually added protection group. I deleted it. Let's see what happens tonight at 9PM

[danswartz]

Looks like I am good to go