Problem with worker VM start initiated by Veeam on PVE

I really like VEEAM B&R with ESXi hosts. But today vmware is under broadcom and has ruined politics, so we need to move on to PVE.

I has same issue on one of three hosts, so i will try every solution, thanks.
 
justanotherITdude said:
I got info from Veeam support and we can actually leave the Veeam worker just in an idle/powered on state.

C:\Program Files\Veeam\Plugins\PVE\Service\
edit the appsettings.json
Under "Workers" you can change KeepTurnedOn from false to true
"KeepTurnedOn": true

Save and reboot server or Veeam PVE Service.

I am going to try this out for awhile and see how it does. It looks to barely use any cpu resources while idle but will eat up some ram but less risky then having it not power up correctly and miss checkpoints on VMs.

I did test what happens if you reboot the Veeam server with worker powered on, the Veeam worker will stay powered on, once a backup is started it resets the Veeam worker and you can see the uptime clock reset.

Seems good so far, will let it go for the rest of the week and see if it's more stable.
Click to expand...


This config works for me, 45 days without issues.
 
In my situation problem was in bad qcow2 file (btrfs filesystem on host storage).

I was unable to migrate it, convert it, etc. QM check don't mate a difference.

I made another one, clone old to it with clonezilla and problem solved.
 
We have also set `KeepTrunedOn` to true. Now i want to change the network settings of my worker, but im not able to because its turned on.
I have tried turning it off in proxmox, but veeam keeps showing it as online.

Is there a way to turn the worker off without having to change the config back and testing the worker?
 
Teletrend AG said:
I had a case opened a while ago too, but it was just closed afer a few days because of "no ressources". I tried to escalate it, but the escalation team basically told the same. I asked why they can't just keep it open until someone has got the time to process it? - the answer was again similar. In short: if you use the Community Edition, don't expect any form of support - they did't even accept it as a bug report...
Click to expand...
have you tried logging into the veeam host and workers and installing she keys. To get into the workers I had to get them into single user mode and create myself an account I also noticed that the firewall on the workers doesn't have ssh enabled, nor is vim or net-tools installed. I am having a different error and I need those for troubleshooting.
Hope that helps.
 
Hello guys,

After another update to Veeam (v13), this issue came back again. (PVE do not start and jobs waiting for a worker)

I followed the same process as below:

C:\Program Files\Veeam\Plugins\PVE\Service\

Edit the appsettings.json file.

Under "Workers", change KeepTurnedOn from false to true

"KeepTurnedOn": true
 
I worked with Veeam advanced support on this issue. We are seeing that qm start task against the worker VM is getting stuck indefinitely and the only way to get the job to run is to kill the task. Veeam provided me with a private build that checks for a stuck task and kills it before trying again and this worked for us. It is not a fix, since the task still gets stuck, and there is no clear reason why. But at least the jobs are successful. According to the case engineer, this update should be included in the next PVE plugin release.
 
  • Like
Reactions: beresa
Hello everyone,

In addition to issuing the aforementioned fix on Veeam's side, we have also filed a proposal to adjust a few more things.

There also remains an open question: why exactly do the VMs hang during startup? If anyone is willing to collaborate and help us and the Proxmox team investigate, please do get in touch.

Thanks!
 
  • Like
Reactions: milenz and Johannes S
It is great to see that this is being escalated to Proxmox. Happy to collaborate if i can be of any help.
 
You must log in or register to reply here.
Top Bottom
Back