Hi Lars,

Sorry it took me a few days to fool with this.  To answer the original
questions: I am not using any NFS, nor mandatory kernel file locking.  I
have experienced this problem with gcc, egcs/glibc1, and pgcc/glibc[12],
all on Linux 2.0.34.  I am also finding this problem with Apache 1.3.1
and 1.3.0, with mod_perl verisions 1.12, 1.14, and 1.15

I tried to reproduce the problem using CGI instead of mod_perl, and I
could not.  When the server dies under mod_perl, it is answering over
100 requests per second.  Under CGI, I simply cannot reproduce that
connection rate, and it loafs along at ~4 requests per second.  This may
indicate that it is a mod_perl problem.

The problem also seems to crop up when I issue 'apachectl stop'. 
Witness this entry in my error_log, when I started the httpd, then
stopped it b/c I stuffed up my code.  The only intervening step was a
mod_perl program which returned 500.

[Tue Jul 21 20:50:40 1998] [notice] Apache/1.3.0 (Unix) mod_perl/1.12
configured -- resuming normal operations
[Tue Jul 21 20:50:44 1998] [emerg] (9)Bad file number: flock: LOCK_EX: 
Error getting accept lock. Exiting!
[Tue Jul 21 20:50:44 1998] [emerg] (9)Bad file number: flock: LOCK_EX: 
Error getting accept lock. Exiting!
[Tue Jul 21 20:50:44 1998] [emerg] (9)Bad file number: flock: LOCK_EX:
Error getting accept lock. Exiting!
[Tue Jul 21 20:50:44 1998] [emerg] (9)Bad file number: flock: LOCK_EX: 
Error getting accept lock. Exiting!
[Tue Jul 21 20:50:44 1998] [emerg] (9)Bad file number: flock: LOCK_EX: 
Error getting accept lock. Exiting!
[Tue Jul 21 20:50:44 1998] [notice] httpd: caught SIGTERM, shutting down

I was hoping to attach gdb to this sucker, but I haven't been able to
get Apache to do this lately, despite my attempts to make its life
difficult.  If I come up with anything I'll get it to you.

Thanks,
Jeffrey

Reply via email to