PMG is mainly for mailserver to mailserver communication. If you are trying sending mail from non mailservers, send those mails to exchange first.
Or
https://www.postfix.org/SMTPD_ACCESS_README.html#relay and test.
Disabled power saving states on the Dell? Firmware updated?
SSDs are enterprise or desktop versions?
I have feeling, you have disk problem, missing virtio drivers in Window VMs etc, but nothing concrete, no VM config/pve versions written.
Maybe...
MTU is 1500 for all servers, switches has 9216. Remotes the same location, the same switch and vlan.
Because internal CA, fingeprints (FP):
Configured for REMOTE1:8006:
from pveproxy-ssl.pem FP - BUT - it automatically detects intermediate...
Nothing spotted yet, because remotes are in production and there is nothing non-standard even on netdata (i can miss something of course).
Today 4x test moving real vm 10+200 GB, migration fails for 2nd disk around 17-19/16-17 GB position...
Created test VM1 with 10 GB and VM2 with 10+10 GB without any OS.
Tested PDM migration:
1st try: VM1 slowed to hell
2nd try: VM2 succeeded
Remotes tests:
SERVER1: if=/dev/pve/vm-104-disk-1 bs=64k status=progress of=/dev/null - skipped due...
The same test as with Alpha version. The same remotes, upgraded to PVE 9. Migration of the 2nd disk stucked, manually stopped.
Remotes: pve-manager/9.0.6/49c767b70aeb6648 (running kernel: 6.14.11-1-pve)
2025-09-19 11:28:53 remote: started...
Filebackup of the host. Especially /etc/pve for pve specific configuration, large part of it is handled by cluster, but something is node dependent (example: certificates).
Search forum at least.
Running PVE 9 on DL360/380 Gen 8-10 without problem.
You can install rsyslog, attach monitor for kernel dump, maybe there will be something not in journal.
It's standard debian anyway.
Other thing - firmware, power, etc etc etc.
Without upgrading the 2nd SAN node:
What will happen, when you shutdown the 2nd SAN node?
What will happen, when you shutdown the 1st SAN node?
I think, it's SAN support thing mainly, because with 2nd SAN node upgrading it blocks something...
That's docker problem versus apparmor, not PVE related. Debian 12 didn't have active apparmor, but debian 13 has it by default. I encountered the same problem with giltab-runner + docker. Profile docker-default is created by docker service on the...
I have SAN with 3 luns connected with multiple destination's paths by multipath without problem. But i used mainly cli for configuration at the end, not gui. Looking on your debug, i think problem is somewhere
1] acl on SAN
2] iscsi/multipath...