Attempt to free unreferenced scalar

[Bill Moseley <mo...@hank.org>]

This message shows its ugly head now and then.

I'm not sure what happens but every once in a while after sending a SIGUSR1
or SIGHUP the "Attempt to free unreferenced scalar" messages start to fill
up the error_log.

The only way to stop them is to really SIGTERM the server and then restart it.

I'm only assuming that this is mod_perl related.  The archives mentioned
that this is probably caused by some problem in a xs module.  If that's
true I can't imagine how to track it down in all the modules that are
running on this mod_perl server.

Does the fact that it takes really killing the server to make it stop and
that a simple restart or graceful tends to start it happening give any
hints to where the problem might be?

BTW - does SIGTERM do a graceful shutdown?  We have been rotating with
SIGUSR1, but then we sometimes find the error_log full of these messages in
the morning.  We could rotate by killing the server and restarting, but I'd
rather gracefully kill the server.

Apache/1.3.9 (Unix) mod_perl/1.21 Perl 5.005_03

Bill Moseley
mailto:moseley@hank.org

Re: Attempt to free unreferenced scalar

[Cliff Rayman <cl...@genwax.com>]

check out these previous posts:
http://www.geocrawler.com/search/?config=182&words=unreferenced+scalar

cliff rayman
genwax.com

Bill Moseley wrote:

> This message shows its ugly head now and then.
>
> I'm not sure what happens but every once in a while after sending a SIGUSR1
> or SIGHUP the "Attempt to free unreferenced scalar" messages start to fill
> up the error_log.
>
> The only way to stop them is to really SIGTERM the server and then restart it.
>
> I'm only assuming that this is mod_perl related.  The archives mentioned
> that this is probably caused by some problem in a xs module.  If that's
> true I can't imagine how to track it down in all the modules that are
> running on this mod_perl server.
>
> Does the fact that it takes really killing the server to make it stop and
> that a simple restart or graceful tends to start it happening give any
> hints to where the problem might be?
>
> BTW - does SIGTERM do a graceful shutdown?  We have been rotating with
> SIGUSR1, but then we sometimes find the error_log full of these messages in
> the morning.  We could rotate by killing the server and restarting, but I'd
> rather gracefully kill the server.
>
> Apache/1.3.9 (Unix) mod_perl/1.21 Perl 5.005_03
>
> Bill Moseley
> mailto:moseley@hank.org