[SOLVED] 2 PMG cluster, both licensed community level, log into either one and it says the other not verified abort!

[bwdavis]

Hi
sf-01 (primary)
sf-02 (secondary)
In cluster.

I just updated and applied patches and upgrades to both servers and restarted them.
Problem:
Logging into to primary server error message:
sf-02:
fingerprint 'FF:FF:FF:FF.....' not verified, abort!
Logging into to secondary server error message:
sf-01:
fingerprint 'AB:CD:FED:FF.....' not verified, abort!

 

[Stoiko Ivanov]

* did you change the certificates on the cluster? (you need to adapt the fingerprints in /etc/pmg/cluster.conf' as well)
* please post the output of:
`pmgcm status`

 

[bwdavis]

* did you change the certificates on the cluster? (you need to adapt the fingerprints in /etc/pmg/cluster.conf' as well)
* please post the output of:
`pmgcm status`

logged in to master node:
pmgcm status
NAME(CID)--------------IPADDRESS----ROLE-STATE---------UPTIME---LOAD----MEM---DISK
sf-01(1) 10.10.11.233 master A 17:00 0.07 61% 14%
sf-02(2) 10.10.11.240 node ERROR: fingerprint 'C2:xx:F3:10:45:8A:xx:42:9A:2A:xx:AD:1D:xx:09:47:xx:xx:B4:A9:8D:xx:91:xx:C1:EA:37:61:3F:3A:xx:19' not verified, abort! - - -% -%
logged into secondary node:
pmgcm status
NAME(CID)--------------IPADDRESS----ROLE-STATE---------UPTIME---LOAD----MEM---DISK
sf-01(1) 10.10.11.233 master ERROR: fingerprint 'br:dm:4F:91:3F:bx:bb:B2:57:96:B0:32:14:xx:2E:9A:xx:04:70:38:5B:8B:d3:mz:0A:61:67:41:53:fa:80:bx' not verified, abort! - - -% -%
sf-02(2) 10.10.11.240 node S 17:06 0.12 57% 13%

 

[tom]

See https://pmg.proxmox.com/pmg-docs/pmg-admin-guide.html#_change_certificate_for_cluster_setups

 

[bwdavis]

OK I put the current keys of the two nodes in /etc/pmg/cluster.conf on both servers, I did the node first, and then the master. The cluster status changed from error to syncing, and now they are both active. The documentation you cite says:
" If you change the API certificate of an active cluster node" you have to update this file with the new fingerprints. How did this occur? AFAIK we didn't change anything on the servers, just noticed the error when logging in. The error is now gone, and the subscription status of both the master and node show normal.