Messages in this thread | | | Date | Mon, 18 Oct 2010 23:31:09 +0800 | From | Wu Fengguang <> | Subject | Re: oom_killer crash linux system |
| |
On Mon, Oct 18, 2010 at 05:10:00PM +0800, KOSAKI Motohiro wrote: > > > > i want to test the oom-killer. My desktop (Dell optiplex 780, i686 > > kernel)have 2GB ram, i turn off the swap partition, and open a huge pdf > > files and applications, and let the system eat huge ram. > > > > in 2.6.35, i can use ram up to 1.75GB, > > > > but in 2.6.36-rc8, i just use to 1.53GB ram , the system come very slow > > and crashed after some minutes , the DiskIO is very busy. i see the > > DiskIO read is up to 8MB/s, write just only 400KB/s, (see by conky).
There are much more reads than writes, it looks like some thrashing. How do you measure the 1.75GB/1.53GB?
> > what change between 2.6.35 to 2.6.36-rc8? is it low performance about > > page reclaim and page writeback in high press ram useage? > > very lots of change ;) > can you please send us your crash log?
And there are several ways to help debug the problem.
- reduce the dirty limit
echo 5 > /proc/sys/vm/dirty_ratio
- enable vmscan trace
mount -t debugfs none /sys/kernel/debug echo 1 > /sys/kernel/debug/tracing/events/vmscan/enable <eat memory and wait for crash> cat /sys/kernel/debug/tracing/trace > trace.log
- log vmstat events
i=1 while true; do cp /proc/vmstat vmstat.$i let i=i+1 sleep 1 done
Thanks, Fengguang
| |