Why doesn't Apache/Stronghold like Linux 2.0.x?

Joshua M. Thompson (invid@optera.com)
Mon, 9 Dec 1996 20:13:06 -0500 (EST)

In the continuing saga of those of us plauged by problems with running
Apache 1.1.1 (or, in my case, Stronghold 1.3.2 -- same thing but with SSL
support), I'm having some damned strange problems since I went up to
2.0.26/27. This problem predates my use of Stronghold (I was having it
back when I ran plain vanilla Apache 1.1.1), so i don't think it's related
to Stronghold.

Originally, i was having the problem many others were having -- I would
end up with httpd processes stuck in the R state, eating up huge amounts
of CPU time. These processes could not be killed; even multiple kill -9 's
would leave them unscathed. I would have to either reboot, or let them die
(which generally took hours).

Now that I've gone up to 2.0.26 (And today 2.0.27, same result) I've got a
stranger problem. Now I ended up with pretty much the same symptoms,
except the processes are killable and they are happening MUCH more often.
Now instead of having this happen every couple days or so, I'm forced to
kill and restart the server every HOUR OR TWO. Suffice it to say it's
stressing me out to no end.

I recall that these problems started way back in the early 2.0.x days,
possibly around 2.0.13. I vaguely recall that some changes were made to
the handling of file locking...

invid@optera.com             | We are Grey
http://www.optera.com/~invid | We stand between the Candle and the Star
                             | Between the Darkness and the Light