[Testbed-admins] Mysqld problem on ops-install

[Armen Babikyan <armenb@ll.mit.edu>]

Hello,

We are in the process of creating a small Emulab cluster.  On my ops 
machine, my Emulab ops-install script gets stuck looking for a 
/tmp/mysqld.sock file.  Upon further investigation, I found the 
/usr/local/etc/rc.d/1.mysql-server.sh file.  When I copy-paste the 
command string into my terminal (appropriately substituting for 
$safeprog), mysqld SIGSEGVs:

ops:~# /usr/local/bin/mysqld_safe --pid-file=/var/db/mysql/mysqld.pid 
--user=mysql --log-long-format --log=/usr/testbed/log/mysql/base 
--log-bin=/usr/testbed/log/mysql/update 
--log-slow-queries=/usr/testbed/log/mysql/slowqueries -O 
max_connections=300 -O long_query_time=2 -O key_buffer=32M -O 
table_cache=256 -O sort_buffer=4M -O record_buffer=1M
Starting mysqld daemon with databases from /var/db/mysql
Segmentation fault
STOPPING server from pid file /var/db/mysql/mysqld.pid
090428 01:09:03  mysqld ended

ops:~# uname -a
FreeBSD ops.foobar.com 6.3-RELEASE FreeBSD 6.3-RELEASE #0: Wed Jan 16 
01:43:02 UTC 2008     
root@palmer.cse.buffalo.edu:/usr/obj/usr/src/sys/SMP  amd64
ops:~#

Since mysqld_safe is a script, I ran it under sh -x to see what the 
actual command that crashed was:

ops:~# /usr/local/libexec/mysqld --basedir=/usr/local 
--datadir=/var/db/mysql --user=mysql 
--pid-file=/var/db/mysql/ops.foobar.com.pid
Segmentation fault
ops:~#

Incidently, although the processor is an amd64, the 32-bit version of 
FreeBSD 6.3 was installed on this machine.

Is this a known issue?  Does anyone have any intuitions about the cause 
of this problem?

Any and all advice is appreciated.  Thanks!

Armen

--
Armen Babikyan
MIT Lincoln Laboratory
armenb@ll.mit.edu . 781-981-1796