but I used BerkeleyDB-4.8 and libevent1.4.13 and memcacheq.0.2.0
No matter what I run to start memcache - it runs - and this outputs:
<23 server listening
<24 send buffer was 129024, now 268435456
[memcacheq] [Fri Jan 22 11:10:48 2010] "qstats dump thread: a qstats
is dump."
[memcacheq] [Fri Jan 22 11:10:48 2010] "mempool_trickle thread:
writing 0 dirty pages"
[memcacheq] [Fri Jan 22 11:10:49 2010] "checkpoint thread: a
txn_checkpoint is done"
but as soon as I telnet 127.0.0.1 22201 or run a $memcache_obj =
memcache_connect('127.0.0.1', 22201);
it dies: Segmentation fault (core dumped)
I have tried a variety of tweaks to the start script, but 100% of the
time, the above occurs.
Anyone have any ideas?
thanks.
Kevin