Project

General

Profile

Actions

Bug #1539

closed

can't capture lighttpd startup notifications output in perl script

Added by Anonymous almost 17 years ago. Updated over 8 years ago.

Status:
Fixed
Priority:
Normal
Category:
core
Target version:
ASK QUESTIONS IN Forums:

Description

When starting lighttpd through a perl script we are trying to work around the problem mentioned in ticket 1452 (no proper exit value). We decided to capture the output and scan it for the text "going down" indicating a fatal startup error came along:


$result=`lighttpd -f /etc/lighttp.conf 2>&1`;
if ( $result =~ m/going down/ ) {
    $RetVal=0;
} else {
    $RetVal=1;
}
return $RetVal;

This code hangs if the startup of lighttpd succeeds. It is visible in the process list that the lighttpd started by the code has gone defunct:br
root@san-radio-1:~# ps -axfbr
2555 ? Ss 0:00 \_ sshd: root@pts/2br
2557 pts/2 Ss 0:00 | \_ -bashbr
2733 pts/2 S+ 0:00 | \_ /usr/bin/perl -w /etc-ro/rc.d/lighttpd.pl startbr
2734 pts/2 Z+ 0:00 | \_ lighttpd <defunct>br
2736 ? S 0:00 /usr/sbin/lighttpd -f /etc/lighttpd.confbr
2737 ? Ss 0:00 \_ /usr/bin/php-cgibr
br

When checking filedescriptors it becomes clear that the perl script has an open filedescriptor on the backtick construct. The defunct process does not show any open fd's

I haven't looked at the code, but it appears to me as an ordering problem in closing the filehandles and sending a signal.

-- mpo

Actions #1

Updated by gstrauss over 8 years ago

  • Description updated (diff)
  • Status changed from New to Fixed
  • Assignee deleted (jan)
  • Target version changed from 1.5.0 to 1.4.40

In lighttpd 1.4.40, proper exit value (if there is an error at startup) is propagated to parent process which starts daemon

Actions

Also available in: Atom