[SOLVED] PBS - how to restart the web GUI? (was: estimate verify runtime)

[wbk]

Hi all,

(Excuse me, this was supposed to go into the PBS subforum)

Is there a way to estimate / extrapolate runtime of backup actions, or guess whether it finished or is still running?

My PBS runs on low end hardware (4TB@7.2k spinning SAS, non-RAID, with LVM cache on SSD, quadcore Atom D525 with 4GB of DDR3, 1Gbit network). I expect backup actions to take a while, and there is no need/incentive to invest in an upgrade. Still it would be nice to have some idea how long to wait before kicking off the next action.

I had a verification task running on a 2.3 TB backup of a container with a handful of snapshots.

The task had been running for 16+ hours when the WebGUI became unresponsive; subsequently I was unable to log in over SSH. A local shell opened, but my first command (`htop`) did not return anything and seems to have caused the system to completely lock up.

This PBS never had this behaviour since I installed it two(?) years ago, so it is not that I suspect the hardware unable to handle the (hardly existent, as limited by disk-IO) load.

Now my question would be:

• Should I hard reset the system, or should I wait?
• If the system does not react to any external stimuli, can I expect the verify task to complete if I leave the server running?

When I last checked, the datastore had a consistent thoughput of some 15 IOps, resulting in 16+ hours of 60 MB/s read on the related disk. That would mean that a lower bound of the total amount of data read would be 50 (MB) x 60 (sec) x 60 (min) x 16 (hour) = 2800000 MB read, devided by 1024 (GB) devided by 1024 (TB) gives at least 2.7 TB of data read. The total amount of data on the disk is 2.37 TB, according to the storage status that is (still) visible in the datacenter/storage/summary view (the last update to that view was an hour ago)

Thanks for any hints, pointers and feedback!


/edit/
The local login came back to live. After that I was able to login via SSH.

When I ask

# proxmox-backup-manager task list

it says:
Error: error trying to connect: error connecting to https://localhost:8007/ - tcp connect error: deadline has elapsed

That is in line with not being able to connect to the WebGUI.

How do I restart proxmox-backup-server?

 

[fabian]

"systemctl restart proxmox-backup-proxy proxmox-backup-daemon"

anything interesting in the system logs? "journalctl --since yesterday", for example..

 

[wbk]

Hi Fabian,

Thanks for your reply and for moving the thread!

Sorry I was not able to reply any sooner. I ended rebooting the server later that day; upon returning, I learned that the manual verification I was waiting for had long since finished, but in the mean time the re-verification of older snapshots had started and was well underway.

By now all is (re-)verified.

"systemctl restart proxmox-backup-proxy proxmox-backup-daemon"

I'll keep that in mind for whenever I might need it in the future.

I will look into the journal later today and post back if I find something that attracts my attention.

Thanks!