Always a first time!
Why not confirm it works with a Backup - or not?
Quick enough to do…
My 2 cents
Andy
1 Like
Also , when using ZFS HA is available . If you use local ZFS, replication is used.
HA is also available when using XFS on a NAS, which IS supported by Proxmox.
I am quite well aware of the capabilities of Proxmox…
Yes, the backups work perfectly. But the snaps are fast and you avoid loading the network on the NIC dedicated to the PBS
What I assumed…
I’m using a dedicated PBS network for some clients, and also a dedicated “Storage” network for shared storage. Example: NS7 VM is 1.4 TB large. No issues with bandwidth or IO!
Backups to PBS and NAS…
Yes, I have also been using PBS for a long time with a nic only for the pbs and the cluster.
The 1.4 TB VM takes about 3 minutes to make a PBS incremental backup…
Sure a snap is faster, but I don’t use snaps…
My PBS does have a connection to the LAN - for updates, but also administration, when needed is via LAN.
The other networks (Cluster, Storage, Backup alll are isolated and have no gateway).
I don’t doubt it! But I don’t know what to say , the only VM that gave problems is NS8 on this config.
I ask you a question : If I force the snapshot locally on the ZFS storage mdi proxmox second is better or not ?
To be honest, I can’t give an answer…
I simply never trusted snapshots enough and have always preferred a full backup. That has always worked.
But why not simply try?
You may have found a bug - or a special case…
My 2 cents
Andy
Yes, it’s already done.
Anyway, thank you for your advice, which I take into account.
1 Like
@Andy_Wismer Last question : in Ns8 as in all servers , I installed the qemu-guest-agent and activated it in the VM profile .
I have the option to disable it and uninstall the package . What do you think?
I’ve always had better experience when using the qemu-guest-agent in all available OSes than without.
But you still can try. Usually you can set some options in the QEMU agent for each specific VM.
I use:
1 Like
Hi I had to edit before 'systemctl --user restart postfix` because it didn’t work 
podman exec -ti postfix postmap /etc/postfix/main.cf.d/sender_rename
Now the emails do not stay in the queue and are rewritten as I wish.
Do you think the change I made is correct ??
Yes, the change is correct. The command was missing so I corrected my post.
1 Like
OK, good. I only have the problem that smartrelay doesn’t do with sasl, (outlook.com mx). On neth7 server I have the same relay . Suggestions?
You could compare the postfix configuration of NS7 and NS8 by using postconf.
1 Like