Busy loop with memcached 1.2.2

[Kari Lavikka]

Hi,

I just upgraded (and downgraded) a few of our memcached servers to 1.2.2 
from 1.2.1. After memory usage reached about 1.4GB every of the memcached 
instances went to busy loop (99.9% cpu and didn't accept any connections).

We have debian linux on 64bit dual opterons with 8GB of memory. Each 
server runs two instances. Libevent is: libevent-1.1a.so.1.0.2

Command line arguments were:
memcached -m 3500 -c 8192 -d -P /var/run/memcached2.pid -p 11212 -u memcached


Some stack traces:

0x0000000000407051 in assoc_find (key=0xc4a1f4 "image_36411271", nkey=14) at assoc.c:503
503             if ((nkey == it->nkey) &&
(gdb) bt
#0  0x0000000000407051 in assoc_find (key=0xc4a1f4 "image_36411271", nkey=14) at assoc.c:503
#1  0x0000000000406a7e in do_item_get_notedeleted (key=0xc4a200 "71", nkey=3707308320, delete_locked=0x0) at items.c:370
#2  0x0000000000404072 in event_handler (fd=12886528, which=2336, arg=0x9a0940) at memcached.c:1086
#3  0x00002aaaaacdc0e2 in event_base_loop () from /usr/lib/libevent-1.1a.so.1
#4  0x00000000004058dc in main (argc=13393, argv=0x7fffffaec028) at memcached.c:2742


0x000000000040704e in assoc_find (key=0x1ce8f74 "relevant_CAMP_FREQ_1490_312_4_2007082014", nkey=40) at assoc.c:503
503     assoc.c: No such file or directory.
(gdb) bt
#0  0x000000000040704e in assoc_find (key=0x1ce8f74 "relevant_CAMP_FREQ_1490_312_4_2007082014", nkey=40) at assoc.c:503
#1  0x0000000000406a7e in do_item_get_notedeleted (key=0x1ce8f98 "2014", nkey=948781420, delete_locked=0x0) at items.c:370
#2  0x0000000000404072 in event_handler (fd=30314392, which=16748, arg=0xdddf10) at memcached.c:1086
#3  0x00002aaaaacdc0e2 in event_base_loop () from /usr/lib/libevent-1.1a.so.1
#4  0x00000000004058dc in main (argc=6838, argv=0x7fffff944ed8) at memcached.c:2742


     |\__/|
     ( oo )    Kari Lavikka - tuner at bdb.fi - (050) 380 3808
__ooO(  )Ooo_______ _____ ___ _ _  _   _    _      _                  _
       ""