Hi all, Hi Mark, I see one change in the newest mod_jk (see below), that stops Apache in my configuration:
The discussion on this bug is valid https://bz.apache.org/bugzilla/show_bug.cgi?id=59184, however I would like to see an explanation o why the shm may fail! Docker 1.12 Debian Stretch with Apache 2.4.23 Mod_jk 1.2.42 compiled from source JK Conf without directives for JkShmFile, JKShmSize (standard values). Apparently Mod_jk 1.2.42 fails at startup if something goes wrong with shared memory. If I revert to 1.2.41 apache starts and I only see the older message in the jk_log. Note that I don't need real loadbalancing, as I have only one worker. If I understood the conf correctly I still have a balancer with 1 worker, right? [Wed Oct 12 16:26:28.326947 2016] [jk:emerg] [pid 7] Initializing shm:/etc/apache2/logs/jk-runtime-status.7 errno=22. Unable to start due to shared memory failure. (errno=22 is Invalid Param in jk_shm_open) [Wed Oct 12 16:26:28.327012 2016] [jk:emerg] [pid 7] Initializing shm:/etc/apache2/logs/jk-runtime-status.7 errno=22. Unable to start due to shared memory failure. (errno=22 is Invalid Param in jk_shm_open) Any idea and hints? Thank You. Best regards Peter 3572<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3572> if ((rc = jk_shm_open(jk_shm_file, jk_shm_size, conf->log)) == 0) { 3573<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3573> apr_pool_cleanup_register(pconf, conf->log, 3574<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3574> jk_cleanup_proc, 3575<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3575> apr_pool_cleanup_null); 3576<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3576> } 3577<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3577> else { 3578<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3578> jk_error_exit(JKLOG_MARK, APLOG_EMERG, s, s->process->pool, 3579<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3579> "Initializing shm:%s errno=%d. Unable to start due to shared memory failure.", 3580<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3580> jk_shm_name(), rc); 3581<http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=markup#l3581> } http://svn.apache.org/viewvc/tomcat/jk/trunk/native/apache-2.0/mod_jk.c?view=diff&r1=1761418&r2=1761419&pathrev=1761419 --- tomcat/jk/trunk/native/apache-2.0/mod_jk.c 2016/09/19 10:18:12 1761418 +++ tomcat/jk/trunk/native/apache-2.0/mod_jk.c 2016/09/19 10:23:46 1761419 @@ -3575,9 +3575,9 @@ apr_pool_cleanup_null); } else { - jk_log(conf->log, JK_LOG_ERROR, - "Initializing shm:%s errno=%d. Load balancing workers will not function properly.", - jk_shm_name(), rc); + jk_error_exit(JKLOG_MARK, APLOG_EMERG, s, s->process->pool, + "Initializing shm:%s errno=%d. Unable to start due to shared memory failure.", + jk_shm_name(), rc); } /* we add the URI->WORKER MAP since workers using AJP14 will feed it */