Regular System Maintenance for NMIS

Like any system some regular maintenance will keep NMIS running smoothly, we have automated as much as we can and will continue to do so, here are a few things to check on. 

Check and Audit the configuration

These commands tell you what folders and permissions are incorrect, and then just fix it.
/usr/local/nmis8/bin/nmis.pl type=audit
/usr/local/nmis8/bin/nmis.pl type=config

Check the NMIS code base for syntax and dependancies.

/usr/local/nmis8/admin/check_nmis_code.pl

Ensure that DNS or host name changes reach the relevant parties

As NMIS mostly runs periodically, out of cron, changes to your infrastructure outside of NMIS are usually picked up automatically. There is, however, one exception: fpingd.pl runs persistently and permanently (to minimize reaction time). NMIS starts fpingd.pl if there is none running and then leaves it to its own devices.

As a consequence, if you make changes to your DNS infrastructure or rename hosts, then it's necessary to notify fpingd.pl of  those changes: simply run /usr/local/nmis8/bin/fpingd.pl restart=true after you've made your changes.

Please note that fpingd.pl caches information heavily, so restarting it very frequently is counterproductive.

Log rotation is the most important factor

Verify logrotate.conf is running in cron and setup for your files.

Cleaning up old RRD files periodically

cd /usr/local/nmis8/database
find . –name "*rrd" -type f -mtime +30 -print -exec rm {} \;
find . -name "*rrd" -type f -size 0 -print -exec rm {} \;

We have added a script to help with this, you can run it from admin folder, it will clean up files in there which are more than 30 days old, e.g.

/usr/local/nmis8/admin/nmis_file_cleanup.sh /usr/local/nmis8 30

Other things

Check the runtime graphs
Check for zombie processes