All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Proxmox KVM VPS freezes 1-2 times a day
I have a KVM VPS with a Romanian provider who promises "30-minute ticket response time guarantee during business hours (8:00 am to 18:00 pm EET Mon to Fri) and 2 hours guarantee during non-business hours".
About two weeks ago my VPS started freezing 1-2 times a day: It is still shown as "running" in their panel but does no longer respond to SSH/Pings and the VNC console stops working. I've already stopped my services on that VPS and moved everything to another, so it shouldn't be a load issue.
Since my ticket with them is open for more than ten days now (to be fair: I got a response about 24h after opening the ticket but had to reopen it as the issue wasn't resolved), I'm wondering if anyone here has seen this before and maybe even has a solution? I don't see any other threads about this particular issue, so I guess it's rather related to my VPS and maybe something I can solve on my own?
Comments
is it from hosting solutions ro?
If yes then tag cociu
And is it a clean install of the VPS or you have stuff on it that may be crashing the VPS?
RAM overprovisioning? Had this issue where if on ZFS, 1/2 of the RAM is automatically allocated to ZFS -- if there's a surge, sometimes VMs can be shut down.
Can you go to the Proxmox host directly?
On most providers (I assume using WHMCS + ModulesGarden), open the NoVNC console, copy the URL and trim it to [IP address]:8006/ and go to the Proxmox panel directly.
Possibly. I can go to their Proxmox panel but I have close to no permissions there so I can't check the node's utilization.
I'm also having this issue lol
Does it "unfreeze" itself after some time or is it completely kaput requiring a restart (or whatever manual magic)? Anything of interest in VMs (esp. kernel) logs?
Can you leave a ping running on the VM to some external IP (of yours) and watch for the incoming ping rate on the (good VM)? Correspondingly, can you also have a ping from a good VM to the suspect VM's IP as well as the host node's IP?
I'm curious to know the network behaviour around the time your VM freezes - it may provide you with a clue on whether it is also something host related (like backups or other IO issues).
I assume this is still a problem - seems a rum one at that and I'm curious as well.
No, it needs to be stopped (shutdown doesn't work) and booted again from their panel.
No, the VM just stops logging.
I did, but the pings just stop. No jitter before the VM freezes.
Yep, it's really strange. I even tried updating the kernel to 5.11 from elrepo just to make sure that I'm not running into some fancy bug. But the VM went down again 5 hours ago...