You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tomcat.apache.org by bu...@apache.org on 2002/04/07 17:02:49 UTC
DO NOT REPLY [Bug 7814] New: -
Sigsegv using mod_jk in Apache2.0 (Worker) under heavy load
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=7814>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND
INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=7814
Sigsegv using mod_jk in Apache2.0 (Worker) under heavy load
Summary: Sigsegv using mod_jk in Apache2.0 (Worker) under heavy
load
Product: Tomcat 4
Version: 4.0.1 Final
Platform: Sun
OS/Version: Solaris
Status: NEW
Severity: Major
Priority: Other
Component: Connector:JK/AJP
AssignedTo: tomcat-dev@jakarta.apache.org
ReportedBy: erich.prchal@siemens.at
I got errors "child pid 19410 exit signal Segmentation fault (11)"
when running mod_jk with ajp13 in a multithreaded worker.
The problem only occurs if the cachesize of the ajp13 worker
is less than the threads opened (or the default 1)
and (therefore) I have traced from jk_endpoint_t::ajp_close_endpoint.
I believe the solution is in jk_mt.h lines 91-108:
#else /* Unix pthreads */
#define JK_INIT_CS(x, rc)\
if(pthread_mutex_init(x, NULL)) rc = JK_FALSE; else rc = JK_TRUE;
#define JK_DELETE_CS(x, rc)\
if(pthread_mutex_lock(x)) rc = JK_FALSE; else rc = JK_TRUE;
#define JK_ENTER_CS(x, rc)\
if(pthread_mutex_unlock(x)) rc = JK_FALSE; else rc = JK_TRUE;
#define JK_LEAVE_CS(x, rc)\
if(pthread_mutex_destroy(x)) rc = JK_FALSE; else rc = JK_TRUE;
#endif /* Unix pthreads */
lock,unlock and destroy seem to be exchanged (it is still the same in native2).
At least on Solaris 8 I no longer have the problem since I changed the
above lines to what would intuitivly be correct.
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>