Hi all: Sorry for the delay. I just realized that this never went out.
We had a quite spirited discussion of dtrace's capabilities and a good crowd of 19 people. As Jim Mauro stated DTrace fills the observability gap. My database is slow, my cpu is 100% used for 3 minutes once a day. WHY? It used to be that stopping the system and dumping a kernel crash dump was the way to debug these sorts of issues. However this is obviously a horrible way to do it. In addition to providing insight to Solaris, dtrace is available for OS X and FreeBSD. Jim started with an overview of what dtrace can do and a demonstration of how it does it magic. Also he discussed some of the design decisions underlying dtrace that mitigate the potential problems of dynamically adding and removing instrumentation from the kernel. After this he proceeded to show a number of examples that could be used to troubleshoot performance problems of a number of issues sysadmins see on a regular basis. We hope to have him back for a followup sometime next year, so if you missed him this time you will get another chance. He has generously made his slides available and we have put them at: http://www.bblisa.org/slides/201207_mauro_dtrace.pdf See all of you in September. -- -- rouilj John Rouillard =========================================================================== My employers don't acknowledge my existence much less my opinions. _______________________________________________ bblisa mailing list [email protected] http://www.bblisa.org/mailman/listinfo/bblisa
