On Mon, 13 Jul 2009, Aleksey Tsalolikhin wrote:

> Hi, David,
>
>  Thanks for your reply.
>
>  First of all, I don't have a /proc/slabinfo.  But I do have
> /sys/slab, not sure what to look at there... 199 directory entries.
>
>  Here is example "vmstat 1" at the time of the problem:
>
> procs -----------memory---------- ---swap-- -----io---- --system--
> -----cpu------
> r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa 
> st
> 2  9  42196  28912    132   1456  884    0  2116    28 2631 2643  0  4 26 69  > 0
> 0  3  42196  27432    176   2780  676    0  3128    20 2185 3149  1  4 17 79  > 0
> 0  5  42204  27956    148   1572  872   44  2028   128 1619 2688  0  3 26 72  > 0
> 0 10  42300  29140    156   2128  884 3088  3044  3108 2674 2431  0  5  8 87  > 0
> 0 14  42232  29412    140   2000 1480 1300  3776  1308 3728 4717  1  5  4 90  > 0
> 0  7  42252  29516    156   1592  884  448  3056   448 1790 1973  0  4 12 84  > 0
> 2 14  42228  28592    156   1196 15032 4992 38060  5088 18954 27756
> 0  3  0 97  0
> 4 44  42232  27124    164   1668 2868 1580  6160  1580 2807 2769  0  3  0 97  > 0
> 1 27  42236  28148    172   2440 2096 1408  7452  1440 2918 2318  0  3  0 97  > 0
> 2 19  42248  28540    156   2252  708  204  4080   204 3182 3390  0  2  0 98  > 0
> 1 17  42296  27620    196   2992  976  228  2808   228 1303 1416  0  1  0 99  > 0
> 1 41  42280  27260    208   2104 3840 1476  7960  1536 4488 4796  0  3  0 97  > 0
> 1 11  42264  27348    216   2752  616   12  2632    72 2672 2492  0  1  3 95  > 0
> 0  6  42260  27924    220   2276  412    0  1840    84 1188 1224  0  1 15 84  > 0

those si/so numbers indicate to me that the box is thrashing the swap. the 
wait time definantly says tht it's spending all it's time waiting for the 
disk.

combined with your other ps numbers this definantly says that something 
was eating up the ram.

unfortunantly figuring out what used up all the memory is beyond my 
expertise.

with a more recent kernel I would say that you should post to the kernel 
mailing list, but with a kernel 8 releases behind (about 2 years of 
development), I think that they would start off by asking you to move to a 
newer kernel. If you post to the kernel list acknowledging this and asking 
for information of what to look for if/when it happens again you should be 
able to get good pointers.

David Lang
_______________________________________________
Tech mailing list
[email protected]
http://lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to