httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Laurent Blume <>
Subject Re: [users@httpd] Apache 2.2.8 hangs on Solaris 9: Deadlock?
Date Tue, 15 Apr 2008 12:32:30 GMT
William A. Rowe, Jr. a écrit :
> Krist van Besien wrote:
>> On Tue, Apr 15, 2008 at 1:42 PM, Laurent Blume <>
>> wrote:
>>>  I'll look for how to change the mutexes I can change. Any other hint
>>>  will be welcome.
>> You could indeed use:
>> AcceptMutex pthread
>> However, your problem might be PHP. Running PHP in a multithreaded
>> apache is not recommended. The reason is that PHP is "glue", that uses
>> a lot of other libraries on your system. It is hard to verify that all
>> the libraries used by PHP are thread safe.
> One option to seriously consider is keeping the worker MPM you desire, but
> to configure PHP through fastcgi (e.g. you could try out the new
> mod_proxy_fcgi, or use an appropriate module).
> This allows your fcgi php processes to be happily single-thread-forked,
> leaving them outside of the server's process space, and still enjoy worker
> mpm for your non-generated content.

Thanks, Krist, William. I knew about the PHP reluctancy regarding
threads, but it had worked for years, so I didn't think about it. It
definitely seems like a plausible culprit. I can at least easily try the
prefork version on one server, and check if I still get error messages
and hangups.

The fastcgi option seems to be an interesting possibility, I'll look
into it as well.

Thanks guys, I'm on my way of a clean prefork Apache rebuild and more

/ Leader de Projet & Communauté    | I'm working, but not speaking for
\ G11N | Bull Services
/ FOSUG          |

The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message