perl-modperl mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug MacEachern <>
Subject Re: exit signal Alarm Clock (14)
Date Thu, 02 Mar 2000 22:47:02 GMT
On Wed, 1 Mar 2000, Bill Moseley wrote:
> That message is generated in http_main.c when a child exists with a signal.


> They seem to always happen about 5:03 minutes after the last logged
> Apache::RegistryNG request for that child.  Apache Timeout is set to five
> minutes.  This leads me to think that some non-mod_perl request is
> following a mod_perl request, and timing out.

when you log the RegistryNG request, to you include the url?  i mean, do
you know which script is triggering the problem?  seeing that would help a
great deal.

> In my testing I'm unable to reproduce the error by doing a request to
> mod_perl script, and then doing a request to a mod_cgi script that sleeps
> for longer than Timeout.  That works fine and as expected.  And I've never
> been able to reproduce the message on my test server.

it's easy to reproduce the error message:

local $SIG{ALRM};
alarm 1;
sleep 2;

SIGALRM restoration is not broken in this case, there's just no handler in
that scope.
we just need to figure out how your code is triggering it.

> I don't want to enable mod_perl 'g' debugging due to the volume on our live
> system.  Is there a way I could just enable this one debugging message?
>         MP_TRACE_g(fprintf(stderr,
>              "mod_perl: restoring SIG%s (%d) handler from: 0x%lx to: 0x%lx\n",
>              my_signame(sigs[i]->signo), (int)sigs[i]->signo,
>              (unsigned long)Perl_rsignal_state(sigs[i]->signo),
>              (unsigned long)sigs[i]->h));
> I tried to remove the MP_TRACE_g() wrapping around fprintf(), but it won't
> compile.  My C skills are weak to say the least.

what's the error message?

View raw message