Mailing-List: contact tomcat-user-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Users List" <tomcat-user@jakarta.apache.org>
Message-ID: <5D83C44941AFD4118B6F0002B302984F4389C2@EXCHANGE_SERVER>
From: Denise Mangano <DeniseM@complusdata.com>
To: 'Tomcat Users List' <tomcat-user@jakarta.apache.org>
Subject: RE: Strange mod_jk error - SOLVED
Date: Tue, 3 Jun 2003 16:09:54 -0400 
MIME-Version: 1.0
Content-Type: text/plain

I was having some Vhost conflicts which I just cleared up (thanks John).  I
ran netstat again, with both Tomcat and Apache running, and these are the
results:
tcp        0      0 *:8009                  *:*                     LISTEN


I was still having the "error opening workers" in apache error log, so I
also removed the path to the workers file from inside the VirtualHost
container to be outside all virtual hosts.  The errors are now gone, and ...
the problem is SOLVED!!!  Phew...

Thank you, Thank you, Thank you :)  That's what a newbie gets for messing
with the configuration.  But thankfully it works now, and it is working the
way I want it to (have a new "small" issue - will post separately).

The last note on this, is that in my mod_jk.log, after starting apache, it
has the following code:
[Tue Jun 03 15:45:14 2003]  [jk_worker.c (118)]: Into wc_close
[Tue Jun 03 15:45:14 2003]  [jk_worker.c (199)]: close_workers got -1
workers to destroy
[Tue Jun 03 15:45:14 2003]  [jk_worker.c (120)]: wc_close, done
[Tue Jun 03 15:45:14 2003]  [jk_uri_worker_map.c (190)]: Into
jk_uri_worker_map_t::uri_worker_map_free
[Tue Jun 03 15:45:14 2003]  [jk_uri_worker_map.c (441)]: Into
jk_uri_worker_map_t::uri_worker_map_close

The weird thing is that it has this code repeatedly, every time ... 12 times
to be exact.  Is that normal?

Thanks again to both of you for all your help :)
Denise

-----Original Message-----
From: PELOQUIN,JEFFREY (HP-Boise,ex1) [mailto:jeffrey.peloquin@hp.com] 
Sent: Tuesday, June 03, 2003 2:24 PM
To: 'Tomcat Users List'
Subject: RE: Strange mod_jk error


You might check in your /etc/services file to see if 8009 corresponds to one
of the names listed in your net stat

you might try the netstat and change the grep to 8009 and see if there is
any mention of it at all.


Given the fact that your original error report was

In jk_endpoint_t::service, Error sd = -1

I would hazard a guess that everything was working originally except that
you had the ajp13 port worngly specified in your workers.properties or in
your server.xml.  From my own tests I can only get anjk_endpoint_t::service
error if apache is able to find the worker but the i/port are wrong in
worker.properties.  You current problem of not being able to find a worker
is most likely an artifact of trying changing things in your configuration
files one too many times.

Suggestion:


1) Stop all instances of apache
2) Stop all instances of tomcat
use kill -9 if need be to ensure no JVMs are left (Unless of course this is
a production server and you cannot be that indiscrimate)

3)Start tomcat

4) telnet tomcat-ajp13-domain tomcat-ajp13-port  (as specified in server.xml
for the ajp13 connector tag)
5) netstat -a | grep LISTEN

one of these tests should demonstrate the existence of your listener.  If
not check catalina.out for ajp13 error messages


6) once you are confident that your listener is correct recreate your
workers.property to have the correct tomcat-ajp13-domain tomcat-ajp13-port
information entered.

7) check that your virtual host is loading this file or place the
JkWorkersFile command outside all virtual host to make sure everyone uses
it.

8) restart apache and watch the mod_jk logs and see that it load your
workers file correctly

9) hit webpage and hopefully it will work.

-----Original Message-----
From: Denise Mangano [mailto:DeniseM@complusdata.com]
Sent: Tuesday, June 03, 2003 11:51 AM
To: 'Tomcat Users List'
Subject: RE: Strange mod_jk error


Ok, dumb question.  Does port 8009 have to be open on my firewall then?  If
so then how was it possibly working before?  I can only telnet on port 80
and port 443 (These are the only ports open on my firewall).

The results of the netstat are:
tcp        0      0 *:32768                 *:*                     LISTEN

tcp        0      0 *:login                 *:*                     LISTEN

tcp        0      0 *:vboxd                 *:*                     LISTEN

tcp        0      0 *:sunrpc                *:*                     LISTEN

tcp        0      0 *:ssh                   *:*                     LISTEN

tcp        0      0 cdiserv:smtp            *:*                     LISTEN

unix  2      [ ACC ]     STREAM     LISTENING     1257   /dev/gpmctl
unix  2      [ ACC ]     STREAM     LISTENING     1324
/tmp/.font-unix/fs7100


I am getting the feeling something is not right... But it did work at one
point with the single virtual host...

In httpd.conf
Port 80
BindAddress *
NameVirtualHost *
<If HAVE_SSL>
  Listen 80
  Listen 443
</IfDefine>
ServerName www (overwritten within Vhosts)

Thanks,
Denise


-----Original Message-----
From: PELOQUIN,JEFFREY (HP-Boise,ex1) [mailto:jeffrey.peloquin@hp.com] 
Sent: Tuesday, June 03, 2003 1:06 PM
To: 'Tomcat Users List'
Subject: RE: Strange mod_jk error


Denise,

I could not see in the email trail if it ever was determined that you truly
have a listener on the ajp13 port for the problem domain

if you do
telnet test.mysite.com 8009 

or whatever the correct domain/port for your current configuration do you
get a valid telnet response?

also if you are on linux, you can do a netstat -a | grep LISTEN to see what
is addresses and port have listeners

Jeff

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org