Search results

If you are getting the error you posted, your nodes are certainly not in sync. Otherwise it would be working. Check and post the output of pvecm status (on all nodes) please.

Your node has fallen off the quorum, i.e. the cluster. As such it is inoperable. Check your cluster networking, make sure the nodes can reach it other (ping, ssh), and see if there's anything in the logs about loosing quorum preceding the entries you posted.

Your PVE version is way outdated. Please try upgrading your system... Personal note: If you're already experiencing issues, it may be a good idea to schedule a maintainence window, make backups of your VMs and just reinstall PVE. If you want a go at fixing it like this, I can't be of much...

Der Reihe nach: Problem ist bekannt, wurde auch schon auf der pve-devel Liste diskutiert. Siehe u.a.: https://lists.proxmox.com/pipermail/pve-devel/2021-October/050267.html https://lists.proxmox.com/pipermail/pve-devel/2021-September/049936.html In welcher config? Damit die firewall...

By default, we ensure that our shipped kernel has protections enabled against all known vulnerabilities. You can always confirm this yourself via the commandline: lscpu | grep "Vulnerability". If you want to further discuss a GUI implementation (I personally don't think that's necessary...)...

Why would you expect node50 to be reset? Did you follow the recommended procedure from our documentation? Note especially this part:

Das liegt wahrscheinlich an overlayfs. Docker verwendet das, um die Ordnerstruktur für die container zu generieren, ohne mehrere Kopien davon zu benötigen. So wies aussieht werden die dann bei deinem Backup als normale files weggespeichert, aber beim restore ist overlayfs natürlich nicht...

SR-IOV devices are detected as normal, seperate PCI NICs (in userspace, for all intents and purposes). So yes, that should be possible, though of course in a VLAN (and not physical port) based configuration, you will be sharing bandwidth.

Unterscheidung: Quell-Storage: Heißt nur so weil Daten draufliegen, muss *nicht* als "storage" in PVE konfiguriert sein, nur irgendwo in der dateisystem hierarchie aufscheinen ("gemounted sein") - das ist da, wo das .vma.lzo liegt Ziel-Storage: Der muss in PVE konfiguriert sein, hier landen die...

Das wäre dann wohl ein VM backup, kein festplatten image. Versuchs mit: qm create <vmid> --archive ./2142424.vma.lzo --storage <store>, wobei 'vmid' eine freie VMID ist, und 'store' ein konfigurierter storage, auf dem die disken der VM am Ende liegen sollen.

This message comes from LVM, it attempts to access these disk entries for some reason and fails. This shouldn't be a hard error though, is there an actual issue during import?

In welchem Format ist das Image? Wo kommt es her?

I believe the limit is whatever your FS gives you at this point. I just attempted to create a (thin-allocated) 30 TB disk on ZFS and it completed fine. Edit: That is to say, I don't believe QEMU applies any limit for it's "raw" block driver at least.

AFAIK "open-vm-tools" is specific to vmware products and will not work with QEMU/KVM as used in PVE. If you experience issues with the QEMU guest agent, it might help to describe those instead. We are aware of some issues with hanging backups, but every information can help us troubleshoot.

Yes. Blacklisting the ahci driver would render *all* your SATA controllers useless for the host (PVE), and specifying the PCI id in vfio-pci.ids would auto-load vfio-pci and you would need to do the unbind/bind dance I posted in reverse to load the ahci driver again. Also, looking at your...

Yes, you need to unbind and rebind the devices manually. The "complex bash script" is really not that bad: #!/bin/sh DEVICE="08:00.0" echo "$DEVICE" > /sys/bus/pci/drivers/ahci/unbind echo "$DEVICE" > /sys/bus/pci/drivers/vfio-pci/bind That should do it, change DEVICE to the PCI address of...

Not entirely sure what you're asking, but here's a shot at it: Routing local (between VMs) and outgoing (internet) traffic via the same interface (as opposed to splitting them between two virtual NICs) is generally fine, unless you're transferring very large amounts of data. The benefit of...

You're not running out of RAM in the VM, the PVE host is running out. Limiting ZFS will not help, as its ARC cache is already shrinking automatically to cope. Can you post the output of a qm list && free command on PVE while both VMs are running?

That's curious. Could you try extracting from the command line as discussed earlier in this thread, and then post your full command line and any error messages/output?

That's an OOM killer log, you're running out of memory on the host. Reduce RAM of the VMs or add more RAM to the physical server.