I’ve been working with a customer for the last month or so. They have implemented Oracle Enterprise Manager and looking to move all their monitoring into Oracle Enterprise Manager . This is a good thing and I’ve been working on getting them moved over. One thing they would like moved into Oracle Enterprise Manager is their Oracle GoldenGate monitoring.
Note: I previously wrote a post on how to configure monitoring for Oracle GoldenGate in OEM12c, check it out here.
After I’ve identified everything, the next thing I needed to do was configure monitoring for Oracle GoldenGate (here). Once configured, I started to receive pending status for all the processes. Normally, I wouldn’t be concern about this because is could take the EM Agent a few minutes to get the information from the JAgent and send it to Oracle Enterprise Manager.
After sometime, these pending statuses started to bug me. After checking the emagent and the jagent, I determined that everything was working and reporting as expected. So what is the problem? I looked in the jagent.log and found the following warning:
WARN Capture – Error in getAttribute: No Object with localId found: :type=capture,name=
Searching MOS, the problem turned out to be a bug (Bug 14244174). This bug is specific to the GoldenGate Plug-In registered instances stuck in a pending status. The bug is also associated with the using Oracle GoldenGate (22.214.171.124.1) with OEM12c (yet another reason to stay on the most current release of Oracle GoldenGate).
Now that I’ve established that this is a bug and that I cannot upgrade until the customer oks the process; how do I get past this for reporting in OEM12c? Fortunately, I found a MOS Note (1487633.1) that will help in working around the issue.
The solution presented in MOS Note 1487633.1 is actually simple, but requires that the specific order outlined in the note. Once all the steps are complete, the Oracle GoldenGate objects will respond with their status.
Note: Being that this is Oracle GoldenGate (126.96.36.199.1), OEM reports that the status for downed extract and replicats are up. This issue is more than likely fixed once upgraded passed 188.8.131.52.10+.
Now, how to resolve this issue? Below are the steps that are outlined in the MOS Note. I have performed theses steps and verified that they worked in resolving the pending status problem.
1. Login to ggsci prompt and stop all processes
2. Remove any initial load/replicats from ggsci prompt by running
info all tasks
delete replicat/extract [name] (only needed if they are initial load extract/replicat)
3. Delete $GG_HOME/dirbdb directory.
4. Delete jagent.log and gsserr.log in $GG_HOME.
5. Login to ggsci prompt and run ‘create datastore’ command
6. Start all proceses, run
7. Login to OEM 12c -> Targets -> GoldenGate.
Set the page refresh rate to 1 minute and wait a couple of minutes, the page should show
the correct status.
Once, I have performed the steps above, within a few minutes I had all the Oracle Golden Gate processes reporting successfully.
Keep in mind, the other issue I had with all down processes reporting up? I get around this once the reporting started to be correct was to place the down processes in blackout mode.