Hi
This is not an answer to your actual problem but it may help you
anyway.
> I have a high-activity Apache web server on a 3.4-STABLE box (cvsup'd a
> few weeks ago) which is spontaneously crashing several times a day
> without logging anything about the crash in the messages file.
STABLE might not be the best choice for a webserver (to all: please
correct me when I am wrong here). I suggest to change to (and stay
with) RELEASE as soon as 4.0 comes out.
As for the absence of logging, this is due to a problem with syslog in
STABLE that was discussed here earlier and is not yet fixed, as far as I
know.
> There is no keyboard and monitor attached, and I am only running some
> very basic software including named, postfix (email, very low traffic)
> and apache (80 processes or so, 30-40 active and the rest usually idle).
You don't want to have 40 idle apache processes. This is a waste of
resources. I suggest setting the config back to the default values (or
slightly above) as they provide a reasonable value for most servers.
> Web transaction rate is 50-80 hits per second. I see some apache (httpd)
> processes exit sometimes with signal 10 or signal 11.
Reading the sig 11 FAQ might help. You can find it here:
http://www.BitWizard.nl/sig11/
> I am suspecting either a hardware problem or an OS bug. It's quite hard
> to completely eliminate the possibility of a hardware problem but the
> machine does have ECC memory, so quite likely the problem is not due to
> bad memory. The machine was stable until subjected to high web hit
> load.
This looks very much like a hardware problem, as a lot of hardware
problems show only under high load. Having ECC memory does hardly decrease
the chance of having bad memory. Again, the sig 11 FAQ might help.
I hope I could be of any help, even though I did not actually provided a
solution to your problem ;)
cheers, Tobe
---
"Never eat anything bigger than your head."
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
