Le Monday 02 May 2005 20:14, Andrew Morton a écrit : > hm. I wonder why you had any cachefs pages anyway. Is the sysrq-P > trace always the same? ok, now after collecting the data after a lot of "sysrq-p" after a lot times running into the same 100% of kswapd0, i am sure: it's always the same trouble. it's also reproducable with rc3-mm3 > Does disabling cachefs in kerel config fix it? it seems so, yes. now i'm running rc3-mm3 with disabled cachefs (config can be found at [1]) for more than 6 hours (also under heavy I/O on harddrives) and it is still behaving normal - no running kswapd0 any more (kswapd0 finally managed to find some sleep ;-) ) greetings, Damir [1] : http://cvs.archlinux.org/cgi-bin/viewcvs.cgi/kernels/kernel26mm/config?rev=1.20&cvsroot=Extra&only_with_tag=CURRENT&content-type=text/vnd.viewcvs-markup -- "Protozoa are small, and bacteria are small, but viruses are smaller than the both put together."
Attachment:
pgp7g0HnzOVlS.pgp
Description: PGP signature
- References:
- 2.6.12-rc3-mm2
- From: Andrew Morton <[email protected]>
- Re: 2.6.12-rc3-mm2 - kswapd0 keeps running
- From: Damir Perisa <[email protected]>
- Re: 2.6.12-rc3-mm2 - kswapd0 keeps running
- From: Andrew Morton <[email protected]>
- 2.6.12-rc3-mm2
- Prev by Date: RE: [patch 1/1] Do not enforce unique IO_APIC_ID for Xeonprocessors in EM64T mode (x86_64)
- Next by Date: RE: [PATCH] Priority Lists for the RT mutex
- Previous by thread: Re: 2.6.12-rc3-mm2 - kswapd0 keeps running
- Next by thread: Re: 2.6.12-rc3-mm2 - kswapd0 keeps running
- Index(es):
 |