After following the steps in the below article to force a resync between the local nas database and the UIM nas tables, the nas alarm processing became very slow (about 1 alarm per 5 seconds).
This is causing the alarm queue to grow when alarms coming in faster than processed.
In the nas.log file, errors similar to the following are observed:
nas: nsLookup ip:<Hostname> => '<FQDN' in xxxxms
If the number of ms is very high, this will cause delays in alarm processing.
Nas probe does a nslookup on the ip address in each alarm. The database.db file has a name_service table to cache previous nslookup results. These must complete before moving on to the next alarm for processing.
The cached entries of the old database.db file can be imported to the new database.db file. This will allow the nslookup refresh to not occur for every alarm.
This will allow the nas probe to use the previously cached nslookup entries and not need to validate the ip address in every alarm.
You can also increase the time to revalidate the name entries from 1h (default) to 1d or more if needed to allow more time between revalidation.
If the time in the log for nslookup is greater than 2-3 seconds (2000-3000ms), there may be issues with nslookup from this server. Please confirm that the dns lookups are responding quickly enough.