Have similar issues.
All VMs that use 0.1.285 will sporadically no longer show their IP but "Guest agent not running".
After a reboot, they work again.
The one VM that is also hosting an MSSQL 2017 application, will also lose the DB connection...
Unfortunately no luck. But since I have I seem to have the same issue as
https://forum.proxmox.com/threads/w2025-virtio-nic-connection-drop-outs.174217/page-2#post-814956
I will continue the discussion there.
interesting would be what storage this vm is on.
the filesystem (seems to be zfs) and the underlying models of the nvme ssds (brand and model and amount).
Not sure how this is relevent. But since you ask; no, I don't believe in clusters for Proxmox :)
HA is IMHO grossly overrated and in many setups it causes more downtime because of added complexity than it actually saves uptime.
Especially...
>So with a local ZFS, and VM RAW disks, you can make use of the native ZFS snapshot feature. Which is (unlike QCOW I believe) instant.
yes, but without fs-freeze from the guest agent and thus most likely having worse consistency
I think to cover some of what's been discussed in the thread as I understand it:
Proxmox's EFI Vars `pre-enroll-keys` option uses EnrollDefaultKeys.efi from the upstream edk2 package...
True. I personally don't use Snapshots from live systems, only shutdown ones. Consistency is one reason, the other one would be that I don't want to store RAM and I even with fs-freeze don't trust my application to handle a rollback gracefully...
I also had my issues with notifications on PVE and PBS and have some open questions still nobody was able to answer.
And apparently, sendmail is legacy.
So that is why I did the following:
- Don't set backup jobs to "Auto". Set them to...
Thank you for making my point :)
I am suggesting that you have storage locally and not over a network.
So with a local ZFS, and VM RAW disks, you can make use of the native ZFS snapshot feature. Which is (unlike QCOW I believe) instant...
I updated some VMs from Windows 10 to 11 and see this error in the event log:
Protokollname: System
Quelle: Microsoft-Windows-TPM-WMI
Datum: 15.11.2025 01:06:21
Ereignis-ID: 1801
Aufgabenkategorie:Keine
Ebene: Fehler...
I am guessing you setup SMTP mail and not sendmail?
You need to set the notification matcher below to send it to SMTP and not sendmail.
So disable the default sendmail and enable your SMTP.
I have no experience with special vdev under Proxmox, only TrueNAS. But I think this won't work well, since you are not working with datasets but zvols.
SLOG is probably also not helping, because your workload is probably not doing any sync...
Maybe :) I am not knowledgeable enough to use the right filters or knowing where to look. That is why was not able to find anything. Do you know where I should be looking? The linked powershell commands in the other thread did not find anything...
I have a problem with only one VM.
That VM for unknown reasons sometimes loses guest agent. This is pretty bad, because that makes also the backup impossible, since I do backups to PBS with stop mode.
Is there a known issue with this config?
I...
So, because I was unable to solve this issue, I set the backup job to use Notification mode: "notification center" and in there I set the a match based on severity with all but info and notice, and there I only use gmail as target.