You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by André Warnier <aw...@ice-sa.com> on 2011/03/22 11:24:54 UTC
Re: [OT] Response Swapping
Mark Thomas wrote:
> On 21/03/2011 23:15, André Warnier wrote:
>> Mark Thomas wrote:
>>> On 21/03/2011 22:06, SQ wrote:
>>>> Mark Thomas wrote:
>>>>> On 18/03/2011 21:05, SQ wrote:
>>>>>> Environment:
>>>>>> Solaris 10
>>>>>> Apache 2.2.16
>>>>>> mod_jk 1.2.31 <---------- Tomcat 5.5.30
>>>>>> JVM 1.6.0_14
>>>>>> Which AJP connector are you using? Is APR/native being used?
>>>>> AJP13.
>>>
>>> That doesn't answer my question. There are multiple AJP connectors
>>> and |
>>> they all implement the AJP1.3 protocol. Exactly which one are you
>>> using? -
>>>
>> Maybe the one above there ?
>
> Nope. That is the web server end. I want to know what is being used on
> the Tomcat end. For 5.5.x there are two options (6.0.x has three, 7.0.x
> has two). It should be one of:
> org.apache.jk.server.JkCoyoteHandler
> org.apache.coyote.ajp.AjpAprProtocol
>
> The start up messages should confirm exactly which is being used.
>
The following is no criticism, and is meant as a constructive comment :
For someone who is /not/ deeply involved in the code of Tomcat, it is not really clear in
the documentation that one can choose between different AJP Connectors at the Tomcat level.
Of course, after your comment above, and now that I have had another look at the on-line
documentation (http://tomcat.apache.org/tomcat-6.0-doc/config/ajp.html e.g.) I also wonder
why this eluded me for so long. All the previous discussions on this list about the
choices between the standard blocking Connector, the Nio Connector and the APR native
Connector *for HTTP* suddenly come to mind now. But in all that time, I never realised
that there was also a similar choice for AJP.
It may be because the only place where it is actually mentioned is on the above page (and
equivalent ones for the other versions), and getting to this particular "protocol" element
requires following this path :
http://tomcat.apache.org/
--> Documentation --> 7.0
--> http://tomcat.apache.org/tomcat-7.0-doc/index.html
--> Configuration
--> Connectors --> AJP
--> Common attributes
--> protocol
If one follows instead what would seem to be a more logical path :
http://tomcat.apache.org/
--> Documentation --> 7.0
--> http://tomcat.apache.org/tomcat-7.0-doc/index.html
User Guide
--> 20) Connectors -->
Table of Contents --> AJP
then one never gets there (no link to http://tomcat.apache.org/tomcat-7.0-doc/config/ajp.html)
Another seemingly logical path would be :
http://tomcat.apache.org/
--> Documentation
--> Tomcat Connectors
but that one never leads there either (also no link).
I also believe that the explanation in the "protocol" attribute could be a bit more clear
regarding the fact that choosing a different "protocol" actually means selecting a
different "Connector class".
It is in fact also pretty hard to find an explanation on the Tomcat site, about what a
Connector really /is/.
Patches still welcome, mm ?
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by Christopher Schultz <ch...@christopherschultz.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
SQ,
On 3/22/2011 12:54 PM, SQ wrote:
> I too didn't realize there was a difference in connectors. Thanks for the
> clarification.
>
> Here is the startup info on the connector:
Yeah, that's not what Mark is looking for. When you first start up
Tomcat, there are some messages that show up in the stdout log. For
instance, if you are NOT using APR yet still have the
AprLifecycleListener enabled, you will see a message like this:
Mar 22, 2011 1:54:24 PM org.apache.catalina.core.AprLifecycleListener init
INFO: The APR based Apache Tomcat Native library which allows optimal
performance in production environments was not found on the
java.library.path: /usr/local/apache-tomcat-6.0.29/server/lib
If you ARE using APR, you'll see a message about the version being used
and it's configured capabilities (ssl, sendfile, etc.).
Oddly enough, I didn't see any startup messages in my stdout log for the
connectors, so I can't show you an example right now.
If you aren't sure, stop Tomcat, delete all your logs, re-start Tomcat
and post the entire logs/catalina.out file (or elsewhere if stdout is
being redirected somewhere else) to the list.
- -chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAk2I5NwACgkQ9CaO5/Lv0PCx4gCeMU4z7C2MMaXg+SJ2rzqWFuqv
s/cAnRxLnJ+CKTpEkX9SC6ByB27oXc0s
=z7MP
-----END PGP SIGNATURE-----
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by SQ <se...@compusystems.com>.
>Yep that is it. You are using the old BIO AJP connector based on the
>classes you are loading. Hmm. Not what I was expecting. It might be
>worth taking a look at this bug:
>https://issues.apache.org/bugzilla/show_bug.cgi?id=50189
>The problem is that we need a test case to investigate this but the
>issue is hard to reproduce. Since you have a test environment where you
>can repeat this, what are the chances of getting a Wireshark trace of
>the AJP traffic for a connection where this happens?
>Mark
>Thanks Mark.
>I looked at the bug and noticed the 5.5 fix will be included in 5.5.34.
Any idea when this comes out? We >were going to try to upgrade anyway, but
maybe we should just wait for 5.5.34.
>I would be happy to supply you with a trace, but right now we're unable to
reproduce. We're trying, but >coming up with nothing. Any
ideas/suggestions on how to reproduce would be great.
We are unable to reproduce the bugs listed above and are unsure how to
proceed at this point. Mark had mentioned that we are using the "old BIO
AJP connector ". Is this something we should change? Anyone have any
suggestions on changes to try or ways to reproduce?
--
View this message in context: http://old.nabble.com/Response-Swapping-tp31185040p31278390.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by SQ <se...@compusystems.com>.
>Yep that is it. You are using the old BIO AJP connector based on the
>classes you are loading. Hmm. Not what I was expecting. It might be
>worth taking a look at this bug:
>https://issues.apache.org/bugzilla/show_bug.cgi?id=50189
>The problem is that we need a test case to investigate this but the
>issue is hard to reproduce. Since you have a test environment where you
>can repeat this, what are the chances of getting a Wireshark trace of
>the AJP traffic for a connection where this happens?
>Mark
Thanks Mark.
I looked at the bug and noticed the 5.5 fix will be included in 5.5.34. Any
idea when this comes out? We were going to try to upgrade anyway, but maybe
we should just wait for 5.5.34.
I would be happy to supply you with a trace, but right now we're unable to
reproduce. We're trying, but coming up with nothing. Any ideas/suggestions
on how to reproduce would be great.
--
View this message in context: http://old.nabble.com/Response-Swapping-tp31185040p31213078.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by Mark Thomas <ma...@apache.org>.
On 22/03/2011 16:54, SQ wrote:
>
> I too didn't realize there was a difference in connectors. Thanks for the
> clarification.
>
> Here is the startup info on the connector:
>
> [Loaded org.apache.catalina.connector.Response from
> file:/usr/local/apache-tomcat-5.5.30/server
> /lib/catalina.jar]
> [Loaded org.apache.jk.core.JkHandler from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
> t-ajp.jar]
> [Loaded org.apache.jk.server.JkCoyoteHandler from
> file:/usr/local/apache-tomcat-5.5.30/server/l
> ib/tomcat-ajp.jar]
> [Loaded org.apache.jk.core.Msg from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomcat-ajp.
> jar]
> [Loaded org.apache.coyote.ActionHook from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
> t-coyote.jar]
> [Loaded org.apache.jk.core.MsgContext from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomc
> at-ajp.jar]
> [Loaded org.apache.jk.server.JkMain from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomcat
> -ajp.jar]
> [Loaded org.apache.jk.core.WorkerEnv from
> file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
> t-ajp.jar]
>
> Please let me know if this is what you were looking for. Sorry for the
> confusion.
Yep that is it. You are using the old BIO AJP connector based on the
classes you are loading. Hmm. Not what I was expecting. It might be
worth taking a look at this bug:
https://issues.apache.org/bugzilla/show_bug.cgi?id=50189
The problem is that we need a test case to investigate this but the
issue is hard to reproduce. Since you have a test environment where you
can repeat this, what are the chances of getting a Wireshark trace of
the AJP traffic for a connection where this happens?
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by SQ <se...@compusystems.com>.
I too didn't realize there was a difference in connectors. Thanks for the
clarification.
Here is the startup info on the connector:
[Loaded org.apache.catalina.connector.Response from
file:/usr/local/apache-tomcat-5.5.30/server
/lib/catalina.jar]
[Loaded org.apache.jk.core.JkHandler from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
t-ajp.jar]
[Loaded org.apache.jk.server.JkCoyoteHandler from
file:/usr/local/apache-tomcat-5.5.30/server/l
ib/tomcat-ajp.jar]
[Loaded org.apache.jk.core.Msg from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomcat-ajp.
jar]
[Loaded org.apache.coyote.ActionHook from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
t-coyote.jar]
[Loaded org.apache.jk.core.MsgContext from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomc
at-ajp.jar]
[Loaded org.apache.jk.server.JkMain from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomcat
-ajp.jar]
[Loaded org.apache.jk.core.WorkerEnv from
file:/usr/local/apache-tomcat-5.5.30/server/lib/tomca
t-ajp.jar]
Please let me know if this is what you were looking for. Sorry for the
confusion.
--
View this message in context: http://old.nabble.com/Response-Swapping-tp31185040p31211930.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
Re: [OT] Response Swapping
Posted by Mark Thomas <ma...@apache.org>.
On 22/03/2011 10:24, André Warnier wrote:
> The following is no criticism, and is meant as a constructive comment :
<snip/>
> Patches still welcome, mm ?
Yep. The offer to help anyone set up an environment so they can create
and test patches still stands.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org