On Wed, 9 Aug 2000, Mark Orr wrote:
Hi Mark,
> I compiled 240test6pre9 last night, and put it thru my
> usual tests. I have a bash script that runs iterations of
> about three dozen commands that dump a lot of output and quits.
> (lots of ps'es, cat'ting big files, some perl scripts, some
> programs that show system status...nothing in the background tho,
> it doesnt run them in parallel -- it's not really a benchmark or
> pure torture test. It's more of an exerciser.)
>
> I find that after 20 or so iterations of the loop, this kernel (pre9)
> really starts to bog down. Normally, there's some disk
> activity on the first pass, but after that not much.
>
> With this kernel it starts out working fine, but by iteration 25
> or so (a minute or two later), it's really dragging...disk is grinding
> away. The commands themselves seem to run a full speed, but the
> loading of commands and shell performance is very sluggish.
What does free say.. what does vmstat output look like what are
your scripts doing etc etc ;-) You need to grind down the list of
possibilities as far as you can.
> Even after you exit out of the script, the system remains slow. Just
> logging in and out and running simple shell commands takes 10 to 20
> seconds (disk buzzing away).
OK, sounds like you may be seeing leakage. If you would like a
copy of IKD (Integrated Kernel Debugging toolkit) for pre9, drop
me a line offlist.. it contains a memory leak detector which can
help track down leaky spots.
> There's no other programs running that could cause that - this
> happens even when running runlevel 1, with nothing but bash and init
> in the ps list, and no kernel modules loaded.
>
> I didnt see this with pre6,7,or 8.
>
> ...not complaining, just reporting.
(keep them coming)
-Mike
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Aug 15 2000 - 21:00:19 EST