[SOLVED] ProxMox DataCenter Notifications Name Resolution error

[thestarswain]

Hello Friends!

Recently have started dabbling in Proxmox after getting bit by the homelab bug and have been really enjoying things!
I was working to setup something new today and have had a hard time finding documentation on how something within Proxmox works - namely cluster notifications.

I currently have 3 proxmox nodes configured in a cluster for management (no HA or anything at the moment). I wanted to setup email alerts for various cluster events and navigated to DataCenter > Notifications.

I added an SMTP Relay as a notification target which uses the email address associated to root@pam for its destination.
This relay point to a Mailjet smtp relay server which I use for my domain.

I put all the authentication information in and went to test email but received the following error:

"Could not test target: could not notify via endpoint(s): SMTP_Relay: Connection error: failed to lookup address information: Temporary failure in name resolution (500)".

Based on the nature of the error I realize/ believe this to be a DNS problem in that the cluster cannot resolve dns for the hostname of the mailjet smtp relay server.
Here is where I am confused though - how do I determine or set the DNS at the cluster level to allow for dns resolution on the notifications?

I have DNS setup on each of my nodes, although not all of the nodes have internet access at the moment (some do). Those that are connected to the internet do not seem to have any dns resolution errors and neither do any of their VMs.
All I can assume is that the cluster is for some reason trying to resolve DNS through one of the offline nodes and then not hopping to an online node in its attempt?

Any help/ information would be greatly appreciated! Thanks in advance!
CH

 

[thestarswain]

Ok, some additional context - I did end up getting this working but Im not entirely sure as to what changed.
As I mentioned - my system is setup in a cluster and some of my nodes have internet access in that they get a static IP on their linux bridge. Others do not have an IP on their linux bridge and as a result do NOT get internet access directly on the node. On those systems though, I am often receiving a dhcp address on the various VMs - the VMs will then get internet access even though the node doesnt have an IP to get routed for access itself.

I rebooted one of the nodes and ensured each of my currently running nodes have ips on their linux bridge - after ensuring both nodes had internet access for dns resolution - I re-tested the setup and received a different error:


"Could not test target: could not notify via endpoint(s): SMTP_Relay: Connection error: error:0A00010B:SSL routines:ssl3_get_record:wrong version number:../ssl/record/ssl3_record.c:354: (500)"

To fix this I ended up having to update the encryption methodology from TLS to STARTTLS - everything is now working like a charm!