500 Internal Server Error ... unexpected end-of-file ... fastcgi process died ..
Added by bleeb over 15 years ago
Today, all of a sudden, half of my smarty-powered sites die with a 500 Internal Server Error - while others seem to work just fine on the same server!
I've tried lighttpd 1.4.20 and 1.4.23, php 5.3.0 and 5.2.11, and am running on Suse 9.3. Already rebooted and power-cycled.
This appears in lighty's log:
2009-10-23 14:44:54: (mod_fastcgi.c.2481) unexpected end-of-file (perhaps the fastcgi process died): pid: 6718 socket: unix:/tmp/php-fastcgi.socket-3 2009-10-23 14:44:54: (mod_fastcgi.c.3299) response not received, request sent: 1113 on socket: unix:/tmp/php-fastcgi.socket-3 for /_blog/index.php , closing connection
and strace -p 6718 shows this:
Process 6718 attached - interrupt to quit
wait4(-1, [{WIFSIGNALED(s) && WTERMSIG(s) == SIGXFSZ}], 0, NULL) = 6732
--- SIGCHLD (Child exited) @ 0 (0) ---
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x40700668) = 6810
wait4(-1, [{WIFSIGNALED(s) && WTERMSIG(s) == SIGXFSZ}], 0, NULL) = 6810
--- SIGCHLD (Child exited) @ 0 (0) ---
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x40700668) = 6848
No clue what's going on here. Any help greatly appreciated!
Replies (1)
RE: 500 Internal Server Error ... unexpected end-of-file ... fastcgi process died .. - Added by bleeb over 15 years ago
Thanks, problem has been solved, unrelated to lighty and quite stupid:
php (up to and including 5.3.0!) dies silently if it's errorlog grows to the size of 2GB.
Since s9y (and smarty) log a GIANT HEAP of "deprecated" messages (which are tricky to get rid of) it made php's errorlog grow to this legacy 2GB limit before the log was rotated.
So php scripts that wouldn't produce log output would still run fine, but if one script had something to log this would kill that php-fcgi instance.