Invalid alarms from snmpcollector

Document ID : KB000005772
Last Modified Date : 18/07/2018
Show Technical Document Details

False alarm or invalid alarm from probes like snmpcollecor, VMware, cdm and others. The threshold configured on the probes don't match the threshold on the alarm screen.

Any environment that you have the probe baseline_engine running.

When you use the admin console interface and enable the check box for the baseline engine at any probe that has the feature baseline available, the probe will start calculate the baseline and will create a threshold based on the information is collected on the baseline period. After that the threshold configuration will be set on the baseline side and not on the original probe side as snmpcollector, VMware, cdm and others. Even if after you uncheck the baseline engine on the probe the alarm sometimes still are getting the threshold based on the B_E and not on the probe.

1. Stop the baseline_engine on the hub where the metric was configured.
2. On the hub, navigate down to the directory where the baseline_engine probe is installed where we are going to edit the actual threshold configuration file. So go to $NIMSOFT/probes/slm/baseline_engine/cache_dir
3. Make a copy of the file – e.g.
4. Although the file has a .zip extension, the format is actually gzip, so open it with 7zip on windows, or uncompress it . with gzip on Linux.
5. Now you should have a file named thresholds -- open this with a text editor.
6. Look for the thresholds with the matching metric ID for the metric that keeps alarming. Just delete the lines that have that metric ID.
7. Save thresholds back to disk
8. Compress thresholds again – On windows, using 7zip, you can simply replace the thresholds text file therein. On linux, use gzip to compress it and name the compressed file
9. Start baseline_engine
10. Verify via the POC that the threshold has been removed – you may also want to validate that existing alarm thresholds/baseline settings are still set (ensuring that the file was not corrupted).

If there was any problem, you can always revert the change: stop baseline_engine, copy back to, and start baseline_engine.