Hi, Lukas -
> Please see commit 7347927a67849a74b67d8b25fb58c033ee79042d on
> git://sourceware.org/git/pcpfans.git lberk/dev
Thanks, nice work! Included some Buffalo folks on cc:, because the
idea for this PR [1] came from their site needs at CCR, to have
job-specific pmlogger data. It would be nice to know whether, with
this facility, a "pmlc one-shot METRIC" widget would be helpful or
redundant.
> Included is a proposed solution for PR1073, monitoring and matching
> the lifetime of a pmlogger instance to a PID specified. I've also
> included qa and have found no regressions when testing the pmlogger
> qa group. Any comments appreciated.
I'd suggest just a couple of things. The man page should show that
the $pid in question must be related to the pmlogger process to the
degree that the kernel must permit a kill($pid,0) to it, which means
in effect that only one's own processes may be monitored. Another
thing is the timing of the checking done in the pmlogger main loop: it
is done apprx. once per alarm-expiry, meaning no more frequently than
the normal metric-polling interval (60s default). The test case looks
nice & simple (though the exact number of recorded entries, 4, might
oscillate to and fro on a busier test machine). (A total tool-shed
nit would be to change "--PID" to "--pid".)
[1] http://oss.sgi.com/bugzilla/show_bug.cgi?id=1073
- FChE
|