Jim's Depository

While updating my systems monitoring I discovered Munin today. Munin captures a wide variety of system information and dumps it into RRD files to ultimately graph it at a central location.

The user interface doesn’t communicate problems well, but it provides the underlying data for you to answer those nagging questions that come up, like “When did our email traffic get so high?” or “Has that disk always run that hot?”

And my install notes:

• Also install sensord, smartmontools, and ethtool when installing the munin-node package.
• Make sure to punch a firewall rule for port 4949 from the central machine. The central machine does not want to get hung on one of the nodes.
• Run sensors-detect after installing sensord.
• Check your syslog after restarting sensord, a bunch of my Dells explode the daemon with a problem in the fan sensor, I have to take lm85 out of their modules. 
• If you are a linux user with SATA drives, then when you install smartmontools, you must edit the /etc/smartd.conf file to comment out the DEVICESCAN line and put in a specific device line with the “-d ata”. May as well add a “-m you@email.address” while you are there so it will notify you. Don’t forget to edit /etc/defaults/smartmontools to enable the daemon.
• In /etc/munin/plugins/ you might wish to take out the ntp* files, unless you care about that. Look where those links go and you will see other plugins you could link in. I added veth* ones for my virtual ethernets and the smart one to get all my drive failure data.
• Again, if on linux with SATA drives you must edit the /etc/munin/plugin-conf.d/munin-node file in a couple of places to tell it to use “-d ata”

Sample bits of /etc/munin/plugin-conf.d/munin-node:

[hddtemp_smartctl] user root env.drives sda env.type_sda ata
[smart_sd*] user root env.smartargs -H -c -l error -l selftest -l selective -d ata