PHP running under mod_fcgid read data timeout in 40 seconds on Plesk

Running a PHP script under mod_fcgid was causing a 500 error to be returned.  The error logged was:

mod_fcgid: read data timeout in 40 seconds
(110)Connection timed out: mod_fcgid: ap_pass_brigade failed in handle_request function

There is a 40 second default on the script running.  To fix this for the requirement we needed to modify the vhosts.conf file for the Plesk domain and put in:

<IfModule mod_fcgid.c>
ProcessLifeTime 7200
IPCCommTimeout  7200
IPCConnectTimeout 300
</IfModule>

And rebuild the configuration file with: /usr/local/psa/admin/bin/websrvmng –reconfigure-vhost –vhost-name=domain

Ubuntu kernel panic twothirdsMD3Transform+0xa3/0x380 SS:ESP

Got an Ubuntu 8.04 (3) system running kernel 2.6.24-24-server which keeps crashing at random intervals.  It can go hours and hours without a crash (even days in some cases) and then other cases it can crash within minutes.  Seeing the following kernel panic (minus some detail):

Seeing:

Call Trace:
add_timer_randomness
scsi_end_request
scsi_io_completion
blk_done_softirq
_do_softirq
irq_exist
do_IRQ
sys_gettimeofday
common_interrupt
out_of_line_wait_on_bit

Code: 0d 89 4c 24 10 01 ce 89 f9 31 d9 21 c1 31 d9 01….

EIP: [<c026f8f3>] twothirdsMD4Transform+0xa3/0x380 SS:ESP 0068:c824be80
Kenerl panic – not syncing: Fatal exception in interrupt

Not sure what the problem is at the moment.  Going to try the 2.6.24-23-server kernel to see if it’s the kernel version, however, I’m wondering if it’s some faulty memory or mainboard.

Poblem was the disks and memory all passed testing and the system was fine for days.

UPDATE: FIXED:

It appears to have been due to some faulty memory modules, although, they tested fine with memtest.  4GB of RAM (4x1GB sticks) were in the machine, removing 2 of these helped and the system hasn’t crashed since.