www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vladislav Malyshkin <...@mail1.nai.net>
Subject Re: mod_jserv/5755: mod_jserv and httpd -HUP restarting problem
Date Mon, 14 Feb 2000 20:20:01 GMT
The following reply was made to PR mod_jserv/5755; it has been noted by GNATS.

From: Vladislav Malyshkin <mal@mail1.nai.net>
To: Ed Korthof <ed@cloudfactory.org>, apbugs@apache.org
Cc:  
Subject: Re: mod_jserv/5755: mod_jserv and httpd -HUP restarting problem
Date: Mon, 14 Feb 2000 14:58:37 -0500

 Why not to get a solution instead of a workarounds.
 As it is now after getting a -HUP signal the other httpd processes
 will try to start JVM machine forever (it was not the case in ApacheJServ 1.0 by the way).
 And httpd will be in some kind of unstable state trying to exec JVM forever.
 -HUP is a standard sygnal used to restart UNIX daemons for ages.
 
 An example how this problem can be fixed is:
 every time before starting JVM ping it first to see if JVM already running.
 If it is already running - stop trying to start JVM from this process.
 You already have all these functions in ApacheJServ-1.1/src/c/*.c , just  call them accordinly.
 
 > Instead of running 'killall -HUP httpd', try sending HUP to only the
 > parent httpd process.  (If you used the RedHat RPM, IIRC this would be:
 > "kill -HUP `cat /var/run/httpd.pid`" -- though you may need to change the
 > file location -- or perhaps something like '/etc/rc.d/apache reload'.)
 
 Many other scripts restart httpd (logrotate and others).
 Also -HUP signal to a daemon may be initiatied by
 many other UNIX programs (in some cases probably even by init).
 
 Vladislav
 
 
 > If
 > you send killall -HUP, the child httpd processes will recieve two HUP
 > signals.  That may (or may not) be sufficient to cause problems of some
 > sort or another with JServ ...
 >
 > Generally, the error which you've described occurs when you're using
 > automatic mode, and mod_jserv fails to tell the JVM to terminate; I'm not
 > sure what could have changed between 1.0 and 1.1, but it's certainly
 > possible that something changed which would affect timing slightly.
 >
 > You can, of course, turn off automatic mode, which would solve this
 > problem -- but then you need to do process management for the JVM
 > yourself.
 >
 > good luck --
 >
 > Ed
 >
 > On 14 Feb 2000, Vladislav Malyshkin wrote:
 >
 > [snip]
 > > I have the following problem
 > >
 > > I  installed ApacheJServ 1.1 which starts automatically by httpd.
 > > It works fine untill first restart of HTTPD by command killall -HUP httpd
 > > Since httpd received -HUP signal every few minutes I get a new message in httpd
server error_log
 > >
 > > ApacheJServ/1.1: Exception creating the server socket: java.net.BindException:
Address already in use
 > >
 > > It looks like httpd is trying to start JVM over and over again despite it is already
running.
 > > In several days I have hundreds of lines with this message
 > [snip]
 >
 > --
 > See you at ApacheCon 2000 in Orlanda, Florida, March 8-10, 2000.
 >                  http://www.Apachecon.com/
 

Mime
View raw message