Re: 99% System load

[Andrew Morton]

Florian Schanda <ma1flfs@xxxxxxxxxx> wrote:
>
> Sometimes (not allways, thats the problem, otherwise I could just avoid the 
>  offending command) during paralel builds (for instance kde-libs or kde-base)
>  the build just stops, and I get a 99% system load.
> 
>  For instance, just now I have (from top):
> 
>  Tasks:  89 total,   5 running,  84 sleeping,   0 stopped,   0 zombie
>   Cpu0 :  0.3% us, 99.7% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
>   Cpu1 :  0.3% us, 99.7% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
>   Cpu2 :  1.3% us, 98.7% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
>   Cpu3 :  1.0% us, 99.0% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
>  Mem:   2071872k total,  1220772k used,   851100k free,    48080k buffers
>  Swap:   401616k total,        0k used,   401616k free,   892428k cached
> 
>    PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
>   5197 ma1flfs   25   0  2256  884 2184 R 99.5  0.0   6:29.38 sh
>   5199 ma1flfs   25   0  2260  888 2184 R 99.5  0.0   6:28.02 sh
>   5201 ma1flfs   25   0  2260  884 2184 R 98.5  0.0   6:29.99 sh
>   5208 ma1flfs   25   0  1560  440 1404 R 98.5  0.0   6:32.03 grep

Make sure that you have enabled CONFIG_KALLSYMS, and that
/proc/sys/kernel/sysrq is set to one.  Then, when it hangs, do

	echo t > /proc/sysrq-trigger

or type alt-sysrq-t

Then, run

	dmesg -s 1000000 > /tmp/foo

and make `foo' available.

It would help if you could make a note of the PIDs of the hung processes,
so they can be correlated with the sysrq output.

Thanks.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/