Problem with ClusterControl server generating increasing numbers of SSH connections to PostgreSQL cluster nodes
I'm seeing my test ClusterControl server generate ever increasing amounts of SSH connections to each of the 3 nodes in my PostgreSQL cluster and eventually, both the ClusterControl server and all the PostgreSQL cluster nodes max out their system memory with SSH connections.
The best interim solution I've found is do a normal full system reboot nightly on the ClusterControl server, but that's not the best solution for moving ClusterControl to a production status from testing and evaluation.
Eric's problem was related to SSSD and SSH configuration on the following line inside /etc/ssh/ssh_config of ClusterControl node:
ProxyCommand /usr/bin/sss_ssh_knownhostsproxy -p %p %h
Commenting that line solved the issue. I'm not sure about your environment (whether you are running SSSD or not) but I hope this will shed some lights for you to troubleshoot the issue.
Please sign in to leave a comment.