www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sunny PARIS <...@weborama.fr>
Subject general/3198: [crit] (98)Address already in use: make_sock: could not bind to port 80
Date Tue, 13 Oct 1998 12:47:45 GMT

>Number:         3198
>Category:       general
>Synopsis:       [crit] (98)Address already in use: make_sock: could not bind to port 80
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    apache
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Tue Oct 13 05:50:01 PDT 1998
>Last-Modified:
>Originator:     sun@weborama.fr
>Organization:
apache
>Release:        1.3.1
>Environment:
Apache 1.3.1 with mod_fastcgi
Linux redhat 4.2 kernel 2.0.33
fcgi.pm 0.42
database msql
>Description:
The problem is quite complex, I do not think it's only apache related.
We have a web server that we kill and restart every nights after doing logs rotation and huge
database operations.
At the end of our cron we kill and restart the httpd. And sometimes it cannot restart.
The problem doesn't reproduce itself automaticaly.
My interpretation is the following : we are using msql wich is single threaded and fastcgi.
We are doing a lot of database manipulation (like a dump), as it is single threaded, fast-cgi's
request timed-out.
I suppose that when resarting, appache try to kill itself, then restart and cannot do it because
of a fast-cgi still waiting (?).
Time out for fcgi is 10s.
There are post about those kind of problems in the bug database, but they are old ones.

How can we do, we really need to have our server up all the time.
I give you a detail of the error log :
[Mon Oct 12 17:45:08 1998] [error] (111)Connection refused: accept: (client socket)
[Tue Oct 13 00:00:01 1998] [warn] child process 3179 still did not exit, sending a SIGTERM
[Tue Oct 13 00:00:02 1998] [crit] (98)Address already in use: make_sock: could not bind to
port 80
[Tue Oct 13 00:00:06 1998] [notice] httpd: caught SIGTERM, shutting down

Question, what is the good method to be sure that apache is up all the time and what is the
nicest way to restart it without having those problems.
Thanks
>How-To-Repeat:
Try to restart an apache server with fast-cgi and send a request to the fast-cgi that you
know it will time out.
Send the request just before killing the httpd.
>Fix:
No :(
>Audit-Trail:
>Unformatted:
[In order for any reply to be added to the PR database, ]
[you need to include <apbugs@Apache.Org> in the Cc line ]
[and leave the subject line UNCHANGED.  This is not done]
[automatically because of the potential for mail loops. ]
[If you do not include this Cc, your reply may be ig-   ]
[nored unless you are responding to an explicit request ]
[from a developer.                                      ]
[Reply only with text; DO NOT SEND ATTACHMENTS!         ]




Mime
View raw message