For the context, we are running our OCS inventory server on version 2.8 with RedHat 7.9 since about 2 years now.
We decided to upgrade to version 2.11 on RedHat 8.6
For this we changed the architecture and are now running an nginx server in front of 2 ocs servers, 1 on 2.8 (primary server), the other on 2.11 (secondary server). The nginx forwards the agent request to both servers. Once it is getting populated enough, we'll switch primary/secondary.
We have observed since a while that some host are disappearing randomly from OCS server 2.8, and then gets registered again, and so on. We hoped that this issue wouldn't happen with 2.11, but since few days this is happening aswell.
This is causing a flow of problem since we have other processes that are based on OCS records to work properly.
Question is, have you ever heard of such issues ?
How can we track it ?
We can't find how and why this is happening.