www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lars Eilebrecht <l...@apache.org>
Subject Re: general/2629: apachectl "fibs" about starting httpd
Date Sat, 18 Jul 1998 19:30:01 GMT
The following reply was made to PR general/2629; it has been noted by GNATS.

From: Lars Eilebrecht <lars@apache.org>
To: (Apache GNATS) <apbugs@apache.org>
Cc:  Subject: Re: general/2629: apachectl "fibs" about starting httpd
Date: Sat, 18 Jul 1998 19:52:27 +0200 (CEST)

 -----Forwarded message <199807172321.AAA16777@ness.csc.liv.ac.uk>-----
 
 From: Richard Lloyd <rkl@csc.liv.ac.uk>
 Subject: Re: general/2629: apachectl "fibs" about starting httpd
 
 > If there is an error in your config file apachectl
 > should display the error message from httpd and the message
 > "apachectl start: httpd could not be started".
 
 Under HP-UX, if you leave the Group setting in httpd.conf as #-1 (which
 is what it's set to by default), you get "httpd started" output by
 apachectl and yet the httpd dies immediately. The error_log says:
 
 [Sat Jul 18 00:11:57 1998] [alert] Child 997 returned a Fatal error... 
 Apache is exiting!
 [Sat Jul 18 00:11:57 1998] [alert] (22)Invalid argument: setgid: unable to set
 group id to Group 4294967295
 
 Similarly giving a non-existent user ID (e.g. #29238 when that UID isn't
 in the password field) for the User field in httpd.conf causes the same
 problem - httpd exits, but apachectl reports "httpd started" - error_log
 in that case says:
 
 [Sat Jul 18 00:13:35 1998] [alert] (2)No such file or directory: getpwuid:
 couldn't determine user name from uid 29238, you pro
 bably need to modify the User directive
 [Sat Jul 18 00:13:36 1998] [alert] Child 1014 returned a Fatal error... 
 Apache is exiting!
 
 > Severity-Changed-From-To: serious-non-critical
 
 I would put it back to serious :-) Problem is that outputting "httpd started"
 is a little misleading in the above cases (although "strictly true" because
 the httpd process does actually start) because it dies during the
 initialisation. A check that the process is still running after a couple of
 seconds would make the message output by apachectl more believable.
 
 I don't know if any other config changes other than User/Group would cause
 httpd to die during initialisation (and after forking, hence the failure
 of the original httpd to return an error status code), but it's something
 apachectl doesn't pick up when it really should.
 
 Richard K. Lloyd,           E-mail: rkl@csc.liv.ac.uk
 Connect,                       WWW: http://www.csc.liv.ac.uk/~rkl/
 3, Brownlow Street,
 Liverpool University,
 Merseyside, UK. L69 3GL
 
 --------------End of forwarded message-------------------------
 

Mime
View raw message