On Monday, 28 May 2012 14:31:40 Kogelheide, Ryan LCTZ:EX wrote: > Apache 2.0.64 using worker/pthreads on Solaris 10 > mod_perl-2.0.4
this is quite old stuff. > I recently installed a new release of an application developed by another > team member. It locked up our apache instances as soon as they came under > load. > > pstack showed that all the Apache HTTPD server processes have all > lwp/threads stuck in > > modperl_tipool_pop > modperl_interp_get > modperl_interp_select > modperl_response_handler_cgi looks like you ran out of perl interpreters. When all configured interpreters are in use every subsequent request that needs an interpreter will stop here. See http://perl.apache.org/docs/2.0/user/config/config.html#Threads_Mode_Specific_Directives > except one that's stuck in > > cond_wait_queue > cond_wait > pthread_cond_wait > pthread_cond_wait > xmlRMutexLock > xmlCatalogCleanup > xmlCleanupParser > XS_XML__LibXML_END > Perl_pp_entersub > Perl_call_sv > modperl_perl_call_list > modperl_perl_call_endav > modperl_perl_destruct This looks like an END{} block out of XML::LibXML being executed when an interpreter is shut down. What it waits for here I don't know. Torsten Förtsch -- Need professional modperl support? Hire me! (http://foertsch.name) Like fantasy? http://kabatinte.net
