Search results

thanks @leesteken, the idea of shrinking the swap partition and using the additional space for the /boot partition got me on the right track :) I didn't use GParted but did it on the running node. In case anyone stumbles on this as well, here's what I did as root (assuming that md0 is mounted...

One of my PVE nodes has been installed some years ago and even though it has always upgraded it to the newest PVE version (currently 8.4.3), one thing has been bothering me for a while. When the node was installed, it was done by using ext4 & mdstat, ending in a partition layout like this: $...

thanks a lot, this is really much appreciated!

thanks again @Fantu for those really helpful explanations! For the time being, I think I have experimented enough with the effects of this likely hardware issue and my next step would be to remove the faulty device from my RAID1 array or "profile" in btrfs terms :) Simply attempting to remove...

First thank you @Fantu and @waltar for your support so far. I've had a little time today to investigate the issue further, and I found some strange things. This time, from the grub command line, which I ended upon reboot, I could successfully do a ls (hd0,gpt3)/ as well as a ls (hd1,gpt3)/...

well, looks like I'm back at square one: the node again only boots into grub. Yesterday evening I decided to start a scrub, and that ended in low level block storage errors like these: [27979.988958] nvme nvme1: I/O tag 604 (b25c) opcode 0x2 (I/O Cmd) QID 8 timeout, aborting req_op:READ(0)...

Thanks for the extensive explanation. What I did with clonezilla is to just see if I could manually mount the two btrfs partitions like I did before with the proxmox live system. And I could, apparently, it just worked. I didn't perform any cloning or whatever, but headed directly to the...

well, and it get's even more fascinating. After just mounting and subsequently unmounting the two partitions in clonezilla, I attempted a simple reboot into proxmox and that worked. After booting, the ring buffer contains heaps of lines like these, however: [Mon Jan 6 15:47:54 2025] BTRFS...

I am afraid, I wasn't clear enough, sorry. blkid shows that both partitions on both NVMes have the same UUID (see my last screenshot). What is different, is the error message I get. One says "device 1 uuid 3f3d..." and when trying with the other NVMe, I get a different "device 2 uuid 6...."...

it's the same, unfortunately, just bailing out with a different UUID

Hi, for my home lab, I wanted to give btrfs a try and since PVE supports btrfs+RAID on root, I decided to give it a try and added a PVE 8.3 node with exactly that into my little 2 node home lab. Things went smoothly so far, but after a recent hard reboot, the node doesn't boot anymore, but...

you are trying to open a binary file (/usr/bin/smartctl) with a text editor, that won't work.

I think it's this commit https://lists.proxmox.com/pipermail/pve-devel/2023-September/058995.html

Thanks @fiona and @fweber for your efforts here, this seems to be a particularly nasty bug ... I can confirm this, too, I could indeed revive such a stuck VM by hibernating and then resuming it: # qm suspend 150 --todisk 1 Logical volume "vm-150-state-suspend-2023-08-03" created. State...

I remember having tried with both KSM and/or ballooning disabled months ago, to no avail. Maybe it improves the situation, but at least for us it did not make a real difference.

Coming back from vacation, I indeed found a stuck Win10 VM. Unfortunately I only noticed it now and I also cannot say how long it has been stuck, because this is an unmonitored VM. The VM doesn't have the proposed mitigations=off configured, because this has come up only during my vacation...

ok, then I will open an support ticket the next time one of our VMs freezes. Not sure how fast it happens, but it will, sooner or later ... I'll keep you posted.

Would it help if you got access to one of the nodes when the freeze occurs? If so, I'd be happy to upgrade that node to Basic support and open an official support request. I guess others would be happy to do the same, because this is really growing into a showstopper.

that's exactly what we're seeing as well. The freezing guests can be tight on memory and we have tried with and without memory ballooning, all to no avail.

At least for us, changing it from "VirtIO SCSI" to "VirtIO SCSI single" has made the biggest difference. We've been seeing far less freezes than before ... but as we all know, only time will tell if this really improves things ... And "VirtIO SCSI single" is actually also the default now.