Scary apache errors moving from prefork to worker

We were helping a client push more data through their apache/tomcat web application running on a Solaris box and decided to switch from the process (prefork) to the threaded(worker) processing model. When we fired up the new apache we started getting errors like these:

[emerg] (45)Deadlock situation detected/avoided: apr_proc_mutex_lock failed. Attempting to shutdown process gracefully.

Not good. After a little digging we found that the default AcceptMutex switched from pthread to fcntl between apache 2.0.49 and 2.0.52.

Adding:

AcceptMutex pthread

to httpd.conf cleared up the errors and we’ve had smooth sailing ever since.

del.icio.us:Scary apache errors moving from prefork to worker digg:Scary apache errors moving from prefork to worker spurl:Scary apache errors moving from prefork to worker wists:Scary apache errors moving from prefork to worker simpy:Scary apache errors moving from prefork to worker newsvine:Scary apache errors moving from prefork to worker blinklist:Scary apache errors moving from prefork to worker furl:Scary apache errors moving from prefork to worker reddit:Scary apache errors moving from prefork to worker fark:Scary apache errors moving from prefork to worker blogmarks:Scary apache errors moving from prefork to worker Y!:Scary apache errors moving from prefork to worker smarking:Scary apache errors moving from prefork to worker magnolia:Scary apache errors moving from prefork to worker segnalo:Scary apache errors moving from prefork to worker gifttagging:Scary apache errors moving from prefork to worker

This entry was posted on Wednesday, March 7th, 2007 at 3:52 pm by steveny and is filed under apache, solaris. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

One Response to “Scary apache errors moving from prefork to worker”

  1. Katherine Hosch Says:
    July 24th, 2007 at 10:34 am

    Thanks so much for this information! I was getting all kinds of these errors after moving to worker mode on the T1000 platform, and they are cleared up.
    BTW, I am running apache 2.2.3.

Leave a Reply