Dear list,
Not that I can think anything can be done about this but--
My system went down after a memory leak (probably X) and a large application
caused the swap to run out. It looked like as soon as the swap space became
critical kswapd thrashed around eating all CPU for ~ 10 minutes preventing me
from logging in as root to kill off either X or the large application-
eventually the system locked solid and power cycling was the only way to go.
After ~ 20 hours uptime.
There is no record in /var/log/messages-
I seem to recall that under k2.2.xx this sort of thing would end up with
something getting a signal -KILL rather than thrashing.
Maybe there should be a swap memory high water mark and kswapd should spot
thrashing resort to some fail-sfae behaviour (maybe it does this already?)
and protect one virtual console or soemthing...
In the mean time I will just keep a closer eye on things.
mtt
Information
uname -a
Linux trouble 2.4.7-10 #1 Thu Sep 6 16:46:36 EDT 2001 i686 unknown
# on a fresh reboot
cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 459632640 443174912 16457728 8192 4988928 88207360
Swap: 945864704 13381632 932483072
Size of "large application" 257Mb
X appeared to have grown to 1Gb (from top so this may not be accurate), but
it is impossible to say as the system was not responsive for that last 10
minutes before it locked.
-
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/
This archive was generated by hypermail 2b29 : Sat Feb 23 2002 - 21:00:31 EST