Working for Joyent, I continue to create lots of technical video (~32 hours of edited material to date). Most recent examples, here, are the start of a series I’m doing with Brendan Gregg (author of the DTrace book). In the course of his research to make these videos, Brendan even made some surprising discoveries about tools that everyone thought they knew all about.

Above: Brendan Gregg discusses vmstat, a performance tool used in Solaris-based operating systems, including Solaris 10, SmartOS, IllumOS, that shows the health of the entire system. Part 2 covers all the fields in detail. Part 3 talks about how vmstat works in Joyent’s SmartOS.

^ Brendan talks about the 1, 5 and 15 minute load averages as reported by tools such as uptime and prstat, and then explains in details how they work on Solaris-based Operating Systems including SmartOS, and reveals why they aren’t really 1, 5 and 15 minute averages.

Below: Brendan discusses the key fields of mpstat, another performance tool used in Solaris-based operating systems, including Solaris 10, SmartOS, Illumos, to shows the health of the CPUs on multi-processor systems. Part 2 covers all the fields in detail. Part 3 is a deeper dive into the fields using other tools.

You might also like:

Joyent Retrospective ZFS Performance Analysis and Tools Building a Business on illumos Brendan Gregg on DTrace