This may not be directly a Slash problem, but it's certainly related to installing Slash.
I'm using Apache 1.3.20+mod_ssl+mod_perl (+mod_php, but that's unrelated, see later).
Naked slash site working okay. In fact, everything looks fine from the browser's perspective on all of my virtual sites, but when I look at the log files, I see segfaults every few minutes. 186 of them over Sunday.
I've changed my log parameters to log PIDs, and tying the segfault (which reports PID) back to the last time that PID was used in a GET can be from a few seconds to well over 10 minutes, and occurs in different pages. It does seem to happen when the site is busy, especially if it's being spidered with WebStripper <sigh>, but the segfault doesn't seem to match the PID of any of the spidered pages. (Although it *could* be faulting before the access is logged, of course.)
Prior to setting up Slash, I've used mod_perl (built without PERL_MARK_WHERE=1) with perl scripts, but never with perl modules. That may be the biggest change... so it might not be caused by Perl code in a module, but by the mod_perl module loading mechanism which I've never used before. It might be a conflict with any other Apache module, I've never taken the time to weed them down to a minimal set. It isn't connected to mod_php, because that was my first suspect, so I rebuilt the server without it with no change in the effects.
All modules are static, mod_so doesn't exist.
I can't find any core files anywhere. (Is there a way to make coredumps happen? I get coredumps when apachectl segfaults, just not when httpd does.)
Probably unrelated, but worth noting: I can't do an apachectl graceful. I get an error that SlashVirtualUser is not a recognized keyword. I have to do a stop/start whenever I make a change.
Thanks for any suggestions.