Hi everyone,
A while ago we migrated our test environment of SCOM 1807 to SCOM 2019. We did this by migrating the database from SQL 2014 to 2016 as well as setting up 2 new management servers on a supported OS (2016 server).
We decommissioned the old SQL server and 2 management servers.
However, since the upgrade our linux agent data collection no longer seems to be working.
Here are the steps that we have taken to migrate the linux monitoring:
- Cross signed the certificates with the tool to the new management servers
- Added the new management servers in the linux resource pool
- Made sure all linux run as profiles were migrated to the linux pool.
- Removed the old management servers from the pool
- Used kevin holman's registry keys to optimize for large environments. (only 11 linux agents active at the moment)
- Installed UR2 and the Linux management packs that come with it.
After the upgrade I have noticed that:
- Windows agents are working
- Network monitoring is working
- No errors or warnings on either management server with the exception of the occassional SNMP error (due to ports that no longer exist)
- Linux performance data is no longer collected
- Monitoring of linux machines in general no longer seem to work
- The verbose logs of the modules on the management servers don't tell me anything.
- I can access the linux servers from the management servers with Putty, with the user that is defined in the runas profiles (account is the same for privileged/action account in test)
- TCP 1270 is open from the management servers as well
I can also no longer install any new linux agent, even when I uninstall the agent manually beforehand (rpm command). Windows agent install fine.
I have been working with SCOM for several years but am sort of at a loss here. We will now be making adjustments to the sudo rules for SCOM 2019, but I don't believe this will make much of a difference since this is only really needed for the elevated commands, which are not needed for performance collection for example.
Clearing the cache on the management servers and restarting the services also has no effect on the monitoring.
Any ideas?
Br,
Jasper