Have a web server that's been acting rather strange lately - seems like
everything is getting killed. I receive a connection refused on 21, 23,
25, 80, 443, etc. Still responds to ICMP, though. Cycle the power,
and the box stays up for another couple of days, only to die again
in this manner. Seems to always happen early morning.
Have a few crumbs of information, from apache and syslog, respectively:
[Fri Sep 17 03:17:52 1999] [notice] child pid 5764 exit signal Bus error (7)
Sep 17 03:17:58 sendmail[5759]: NOQUEUE: SYSERR(root): SMTP-MAIL: died on
signal 9
I'm guessing all of my network services are getting clobbered like this...
Would a heavily overloaded system cause errors like this, or is it
some kind of hardware issue?
Sys is a Celery 400MHz (no o/c), 160MB of PC66 SDRAM (memory bus is @ 66MHz,
so this shouldn't matter, right?), IDE disk. 2.2.11 kernel.
Have another machine almost identical to this box (same mobo, cpu,
slightly less memory), with 125 days of uptime on it - hasn't crashed once.
2.2.7 kernel.
Any ideas out there?
TIA
--rgiles
---------------------------------------------------------------------------
Send administrative requests to [EMAIL PROTECTED]
