We are having a problem with Apache (1.3.9) crashing at random intervals using Red Hat 6.1 (Kernel 2.2.1-20smp) with dual processors. Apache will crash anywhere from once to 20 times a day. Sometimes just the parent Apache process dies or locks up, and sometimes all the httpd processes die. There is no pattern. There do not seem to be any clues or anomolous messages in the Apache or system log files. Everything else seems to work just fine (But then again we are not doing much else with the system). Our system consists of an ASUS P2B-DS Motherboard with dual Intel PII-400 cpu's with onboard Adapted AIC-7890AB SCSI Controller, and two Intel Etherexpress Pro 10/100 Ethernet cards. I've tried reinstalling the OS (Both Red Hat 6.0 and 6.1), reinstalling and recompiling Apache (1.3.6 and 1.3.9 with PHP), and checking every hardware and software setting and nothing seems to stop the random crashing. The strange thing is that Apache seems to crash more frequently in the middle of the night when there is little or no traffic on our site. It almost seems that the smaller the http load, the more frequently it crashes. I finally solved the problem by removing one of the PII-400 cpu's from the motherboard. Is there something wrong with the SMP Kernel?
Doug, can you offer any opinions or insights into this one?
With the second processor installed, start apace, then run strace -p <pid of parent apache process> and capture the output of that command to disk. Try to do this at night when apache is likely to die. Wait for it to die, then send the last few hundred lines of that output here. Note: if apache takes very long to die, this could be a *large* file. You may need to stop and restart the strace program a few times, truncating the output file each time to keep it from getting too large.
closed due to lack of feedback and no other reports of similar nature.