On Wed, Oct 20, 2010 at 09:36:47AM +0800, Figo.zhang wrote:
quoted text
> >
> > >
> > > echo 5 > /proc/sys/vm/dirty_ratio
> >
> > reduce the dirty_ratio, i can use memory up to 1.75GB, and then it
> will
> > call oom-killer.
> >So it helps.  Are there intensive IO after reducing 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
> > would you like to help to see the trace.log , i add attached file.
> > There are many vmscan writes showing up in the trace.
> 
> yes, the DiskIO is be not aggressive when i reduce the dirty_ratio, but
> when i use memory up to 1.75GB (total 2GB), the system suddenly crashed.

You seem to use the term "crash" for both "OOM-killed" and "kernel
panic". You mean .36 kernel will panic on memory pressure while .35
kernel will OOM kill the Xorg task?

quoted text
> when i reboot and want to see the /var/log/message, it is without useful
> information.
> 
> so it is other useful debug approach to find the issue?

Documentation/networking/netconsole.txt

eg. netconsole=@:/eth0,6666@10.239.51.110/00:30:48:fe:19:94

You'll need another machine to catch the panic log.

Thanks,
Fengguang
--
unsubscribe notice
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/