dmesg, gdb and output could help

2011/7/13  <[email protected]>:
>
> Comment #16 on issue 99 by [email protected]: Memcached 1.4.2 server
> segmentation fault
> http://code.google.com/p/memcached/issues/detail?id=99
>
> Hi,
>
> Thank you for checking in, and yes, I am still watching this ticket as it
> still continues to be an issue for me.  I manually apply my patch on every
> upgrade to eliminate the "Segmentation fault".  The patched memcached runs
> indefinitely until it is manually restarted after a version upgrade (months
> at a time).
>
> Here is the information you requested:
>
> memcached: 1.4.6-rc1
> OS: linux 32 bit
> libevent: 2.0.12
> gcc: 4.4.4-r2
> kernel: 2.6.30-r8
>
> I did a default build of memcached from the source downloaded from the
> memcached repository.  I was able to get a SIGSEGV from a single request
> using the same test script included with the original bug filing (using a
> new size of 524288):
>
> ./memcached -l localhost -vvv
> slab class   1: chunk size        80 perslab   13107
> slab class   2: chunk size       104 perslab   10082
> slab class   3: chunk size       136 perslab    7710
> slab class   4: chunk size       176 perslab    5957
> slab class   5: chunk size       224 perslab    4681
> slab class   6: chunk size       280 perslab    3744
> slab class   7: chunk size       352 perslab    2978
> slab class   8: chunk size       440 perslab    2383
> slab class   9: chunk size       552 perslab    1899
> slab class  10: chunk size       696 perslab    1506
> slab class  11: chunk size       872 perslab    1202
> slab class  12: chunk size      1096 perslab     956
> slab class  13: chunk size      1376 perslab     762
> slab class  14: chunk size      1720 perslab     609
> slab class  15: chunk size      2152 perslab     487
> slab class  16: chunk size      2696 perslab     388
> slab class  17: chunk size      3376 perslab     310
> slab class  18: chunk size      4224 perslab     248
> slab class  19: chunk size      5280 perslab     198
> slab class  20: chunk size      6600 perslab     158
> slab class  21: chunk size      8256 perslab     127
> slab class  22: chunk size     10320 perslab     101
> slab class  23: chunk size     12904 perslab      81
> slab class  24: chunk size     16136 perslab      64
> slab class  25: chunk size     20176 perslab      51
> slab class  26: chunk size     25224 perslab      41
> slab class  27: chunk size     31536 perslab      33
> slab class  28: chunk size     39424 perslab      26
> slab class  29: chunk size     49280 perslab      21
> slab class  30: chunk size     61600 perslab      17
> slab class  31: chunk size     77000 perslab      13
> slab class  32: chunk size     96256 perslab      10
> slab class  33: chunk size    120320 perslab       8
> slab class  34: chunk size    150400 perslab       6
> slab class  35: chunk size    188000 perslab       5
> slab class  36: chunk size    235000 perslab       4
> slab class  37: chunk size    293752 perslab       3
> slab class  38: chunk size    367192 perslab       2
> slab class  39: chunk size    458992 perslab       2
> slab class  40: chunk size    573744 perslab       1
> slab class  41: chunk size    717184 perslab       1
> slab class  42: chunk size   1048576 perslab       1
> <31 server listening (auto-negotiate)
> <32 send buffer was 107520, now 268435456
> <32 server listening (udp)
> <32 server listening (udp)
> <32 server listening (udp)
> <32 server listening (udp)
> <33 new auto-negotiating client connection
> 33: going from conn_new_cmd to conn_waiting
> 33: going from conn_waiting to conn_read
> 33: going from conn_read to conn_parse_cmd
> 33: Client using the ascii protocol
> <33 set test 0 10 524288
> 33: going from conn_parse_cmd to conn_nread
>>
>> NOT FOUND test
>> 33 STORED
>
> 33: going from conn_nread to conn_write
> 33: going from conn_write to conn_new_cmd
> 33: going from conn_new_cmd to conn_waiting
> 33: going from conn_waiting to conn_read
> <34 new auto-negotiating client connection
> 33: going from conn_read to conn_closing
> <33 connection closed.
> 34: going from conn_new_cmd to conn_waiting
> 34: going from conn_waiting to conn_read
> 34: going from conn_read to conn_parse_cmd
> 34: Client using the ascii protocol
> <34 get test
>>
>> FOUND KEY test
>> 34 sending key test
>> 34 END
>
> 34: going from conn_parse_cmd to conn_mwrite
> Segmentation fault
>
> ---
>
> After doing a clean build using the same patch file I provided you, the
> daemon seems to be stable again (100 successful requests using the same test
> script).  Thanks for looking into this once more, and let me know if you
> need additional information.
>
> thom
>
>



-- 
Roberto Spadim
Spadim Technology / SPAEmpresarial

Reply via email to