The latest upload (1.29.0.2-4) with the revised patch seems to work just
fine. Thanks guys.
Dave
On Wed, 10 Mar 2004, Don Armstrong wrote:
> I haven't seen any segfaults involving this patch on the setup I've
> got...
>
> Dave: Can you possibly build libapache-mod-perl with debugging s
Looks like HTML::Mason::ApacheHandler is what is being parsed when the
segfault happens. That module does reference dir_config() several times,
as shown here:
351:my $config = $r->dir_config;
423:my $config = $r ? $r->dir_config : Apache->server->dir_config;
548:@val = HAS_T
'.
Thanks
Dave O
On Sun, 29 Feb 2004, Matthew Wilcox wrote:
> On Sun, Feb 29, 2004 at 05:37:08AM -0600, Dave O wrote:
> >
> > Ok, it looks like rotatelogs (which is what I meant in my previous email)
> > no longer comes with apache-common? In that case, it was apache
> > att
Ok, it looks like rotatelogs (which is what I meant in my previous email)
no longer comes with apache-common? In that case, it was apache
attempting to open a pipe to a nonexistant program to log which hangs.
Still think this is a bug :-)
I once again tried to upgrade to the latest apache (now 1.3.29.0.1-5) and
still found this behavior. However, I managed to discover that the thing
that was hanging the apache processes was actually logrotate stuck in a
write() call that never returns. This behavior never happened with apache
1.3
I again tried to upgrade to 1.3.29 and this is still happening.
I tried setting MaxClients down to 50 so as not to use a ridiculous number
of resources, and found that after all 50 children had been spawned and
used, no pages could be served. Browsers could connect, but just waited
for the reque
7 matches
Mail list logo
