Re: 2.6.14 kswapd eating too much CPU

[Jan Kasprzak <kas@fi.muni.cz>]

Nick Piggin wrote:
: Can't see anything yet. Sysrq-M would be good. cat /proc/zoneinfo gets you
: most of the way there though.
: 
: A couple of samples would be handy, especially from /proc/vmstat.
: 
: cat /proc/vmstat > vmstat.out ; sleep 10 ; cat /proc/vmstat >> vmstat.out
: 
: The same for /proc/zoneinfo would be a good idea as well.

	I will send it tomorrow - I will try to boot 2.6.15-rc2
to see if the problem is still there.

: Also - when you say "too much cpu time", what does this mean? Does
: performance noticably drop compared with 2.6.13 performing the same cron
: job? Because kswapd is supposed to unburden other processes from page
: reclaim work, so if it is working *properly*, then the more CPU it uses
: the better.

	As you can see from the graphs at
http://www.linux.cz/stats/mrtg-rrd/cpu.html
- the total CPU usage in the new kernel/with more memory is significantly
higher. I have not measured the time of the cron job itself, just the total
system utilization.

-Yenya

-- 
| Jan "Yenya" Kasprzak  <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839      Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/    Journal: http://www.fi.muni.cz/~kas/blog/ |
> Specs are a basis for _talking_about_ things. But they are _not_ a basis <
> for implementing software.                              --Linus Torvalds <
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/