Jim, James.. My thanks for the info! The fog is starting to lift.
It is that averaging from iostat and other utilities that has been making tracking down the actual activities taking place that has forced me to DTrace. One last question about the graph in Figure 4.7. I understand the bit about dealing with start/end times to figure out length of time. I did that in my spreadsheet. The part of the graph that is confusing me still is the offset of the 'strategy' line from 0 (zero) time. What is involved in figuring out what that ~100us offset from 0 for a bunch of block addresses? Where would THAT time offset come from with the DTrace script? Take care. pdc -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Sunday, November 30, 2008 11:35 PM To: Jim Mauro Cc: Clayton, Paul D; dtrace-discuss@opensolaris.org Subject: Re: [dtrace-discuss] Concurrent Disk Activity from "Solaris Performance & Tools" Chap. 4.17.4 Keep in mind that the service times are *averages* over the interval. The average for the data in his spreadsheet would be pretty nice but would obscure the two pretty bad times he saw... Jim --- Jim Mauro wrote: > Hey Paul - I should add that "iostat -xnz 1" is a great method > for determine how well the SAN is performance. > The asvc_t times are disk IO service times in milliseconds. > I usually start there to sanity check disk IO times... > > Thanks, > /jim _______________________________________________ dtrace-discuss mailing list dtrace-discuss@opensolaris.org