RE: URGENT jk_connect.c problem / UPDATE: weird log entry in mod_ jk.log & other problems

[GOMEZ Henri <hg...@slib.fr>]

>i have some strange entries in my mod_jk.log, generated almost 
>every minute:
>
>[jk_connect.c (143)]: jk_open_socket, connect() failed errno = 111
>[jk_ajp12_worker.c (152)]: In jk_endpoint_t::service, Error sd = -1
>[jk_uri_worker_map.c (335)]: 
>jk_uri_worker_map_t::uri_worker_map_close, NULL parameter
>[jk_uri_worker_map.c (185)]: In 
>jk_uri_worker_map_t::uri_worker_map_free, NULL parameters
>
>does anyone here have an idea what the problem could be?? we 
>don't have these errors on the test server, but the production 
>machine generates them every minute. i don't see anything that 
>would misbehave, but it sure scares me not to know what is 
>happening here.

Could you send us a copy of mod_jk.conf and workers.properties


>>in addition, tomcat occasionally dies with a 
>>
>># An unexpected exception has been detected in native code 
>outside the VM.
>># Program counter=0xblahblah (don't have the exact code anymore)
>>#
>># Problematic Thread: prio=1 tid=0xblaah nid=0xblah runnable 
>(don't have the exact codes anymore)
>>#
>>Abort
>
>this is fixed, it was a strange problem with oro-regex that we 
>can reproduce only on the production server (unfortunately). 
>when oro (perl5 matcher) was fed with a control character 
>(\u0000-\u001f) it died and took tomcat with it (stack overflow). bad!

A JVM problem, not a Java problem

>i'm running a 2.2.16 linux kernel, 4-cpu box, tomcat 3.2.2, 
>apache 1.3.19, mysql Ver 11.15 Distrib 3.23.38, oro-2.0.2, 
>jdk1.3.0_02,....

May I recommand you use the latest JDK 1.3.1 or IBM SDK 1.3
(which is very stable under Linux boxes)