[SOLVED] Proxmox Backup Server stops responding

[MyastanPatrin]

Hello all, I’m planning on having a backup system for my home PVE with a local PBS and a off-site PBS over a VPN.

But sometimes both stop responding (not at the same time) is there any way I can try and find the reason? Heat? Resources exhausted?

Thanks in advance!

 

[dcsapak]

what exactly do you mean with 'stop responding' can you access them by any means (e.g. a local console/ssh/etc)
did you already check the syslog/journal (e.g. with 'journalctl') ?

 

[MyastanPatrin]

what exactly do you mean with 'stop responding' can you access them by any means (e.g. a local console/ssh/etc)
did you already check the syslog/journal (e.g. with 'journalctl') ?

Hello! Thanks for your quick response!

No SSH is available, the fans are working but the machine seems inaccessible.
I did not had a look into journalctl yet. Anything special i should look into?

Thanks again!

 

[MyastanPatrin]

I realized that after a while the system recovers without restarting. Very strange

 

[Dunuin]

Also have a look at the IO delay graphs. If the PVE/PBS installation shares the disk with the guests/datastore and you use some cheap storage like SMR HDDs, then the workload can increase the IO delay that high that PVE/PBS isn't responding anymore because it has to wait for the disks too.

 

[MyastanPatrin]

Also have a look at the IO delay graphs. If the PVE/PBS installation shares the disk with the guests/datastore and you use some cheap storage like SMR HDDs, then the workload can increase the IO delay that high that PVE/PBS isn't responding anymore because it has to wait for the disks too.

I can confirm that PBS use the same disk as the datastore. I'm using a mac mini. If IO delay is the cause, Is it possible that using traffic control resolves this issue?

 

[Dunuin]

So what do your IO delay graphs display when the hosts are under heavy load? you should have a look at that graph right after the webUI is working again.

 

[MyastanPatrin]

After searching a bit more, it seems I'm suffering from this issue

https://bugzilla.kernel.org/show_bug.cgi?id=12877