You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tomcat.apache.org by Mark Thomas <ma...@apache.org> on 2012/11/15 22:16:53 UTC
Tomcat 7.0.33 progress report
The 7.0.33 release is taking longer then planned as my pre-tag checks
are still ongoing.
The short version is that the Servlet TCK is failing with httpd +
mod_proxy_ajp but not with httpd + mod_jk and I am still investigating.
The long version is that I have taken a quick look at the failures but
it appears that the reported failures are symptoms rather than the root
cause (i.e. another test is getting Tomcat's internals into a bad state
and the next test that re-uses whichever object it is that is in a bad
state fails). Tracking down the test(s) triggering the problem will take
time. Since mod_jk is OK, it could be a mod_proxy_ajp bug we have hit
with the changes to the Async error handling. I have just updated the
httpd version I use for the TCKs and am re-running the TCK.
I'll provide an updated progress report in a few days or just create the
tag if I resolve the issue.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: Tomcat 7.0.33 progress report
Posted by Mark Thomas <ma...@apache.org>.
On 16/11/2012 17:02, Mark Thomas wrote:
> I'll shout if I get stuck although the fastest way to find my errors
> seems to be writing an e-mail to the dev list as shortly after I send
> it, I spot the silly mistake that invalidates most of what I just wrote.
As it was in this case.
The TCKs all passed as well as the unit tests but the unit test failure
in trunk indicated an error that existed in 7.0.x and trunk. This has
been fixed and although it shouldn't impact the results of the unit
tests or the TCKs I am re-running them before I upload the release
candidate for voting.
We should have an RC for voting ready by Monday at the latest.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: Tomcat 7.0.33 progress report
Posted by Mark Thomas <ma...@apache.org>.
On 16/11/2012 16:51, Rainer Jung wrote:
> On 16.11.2012 11:18, Mark Thomas wrote:
>> On 16/11/2012 04:03, Konstantin Kolinko wrote:
>>> 2012/11/16 Mark Thomas <ma...@apache.org>:
>>>> The 7.0.33 release is taking longer then planned as my pre-tag checks
>>>> are still ongoing.
>>>>
>>>> The short version is that the Servlet TCK is failing with httpd +
>>>> mod_proxy_ajp but not with httpd + mod_jk and I am still investigating.
>>>>
>>>> The long version is that I have taken a quick look at the failures but
>>>> it appears that the reported failures are symptoms rather than the root
>>>> cause (i.e. another test is getting Tomcat's internals into a bad state
>>>> and the next test that re-uses whichever object it is that is in a bad
>>>> state fails). Tracking down the test(s) triggering the problem will
>>>> take
>>>> time. Since mod_jk is OK, it could be a mod_proxy_ajp bug we have hit
>>>> with the changes to the Async error handling. I have just updated the
>>>> httpd version I use for the TCKs and am re-running the TCK.
>>>>
>>>> I'll provide an updated progress report in a few days or just create
>>>> the
>>>> tag if I resolve the issue.
>>>
>>> 1. I wonder whether configuring the logging to include a thread name
>>> will help here
>>> ( org.apache.juli.OneLineFormatter and %I in AccessLogValve)
>>
>> Probably not. Past experience is that it is the Processor that needs to
>> be tracked. There should be a way to get that logged.
>>
>> As it happens some further digging showed that the failure did not
>> depend on any other tests (I missed this the first time around as I was
>> testing in a different environment where the test passed). Now I have a
>> repeatable failure in a relatively simple test case I can start looking
>> at it more closely. That is today's task.
>
> If I can help around ajp let me know and send some details.
Thanks. I am making progress now. A couple of configuration errors on my
part (changed wrong connector in server.xml to NIO) sent me down a few
blind alleys. It looks like Tomcat was doing the wrong thing (not
sending an end message and mod_jk and mod_proxy_ajp were both getting
upset but in different ways.
I'll shout if I get stuck although the fastest way to find my errors
seems to be writing an e-mail to the dev list as shortly after I send
it, I spot the silly mistake that invalidates most of what I just wrote.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: Tomcat 7.0.33 progress report
Posted by Rainer Jung <ra...@kippdata.de>.
On 16.11.2012 11:18, Mark Thomas wrote:
> On 16/11/2012 04:03, Konstantin Kolinko wrote:
>> 2012/11/16 Mark Thomas <ma...@apache.org>:
>>> The 7.0.33 release is taking longer then planned as my pre-tag checks
>>> are still ongoing.
>>>
>>> The short version is that the Servlet TCK is failing with httpd +
>>> mod_proxy_ajp but not with httpd + mod_jk and I am still investigating.
>>>
>>> The long version is that I have taken a quick look at the failures but
>>> it appears that the reported failures are symptoms rather than the root
>>> cause (i.e. another test is getting Tomcat's internals into a bad state
>>> and the next test that re-uses whichever object it is that is in a bad
>>> state fails). Tracking down the test(s) triggering the problem will take
>>> time. Since mod_jk is OK, it could be a mod_proxy_ajp bug we have hit
>>> with the changes to the Async error handling. I have just updated the
>>> httpd version I use for the TCKs and am re-running the TCK.
>>>
>>> I'll provide an updated progress report in a few days or just create the
>>> tag if I resolve the issue.
>>
>> 1. I wonder whether configuring the logging to include a thread name
>> will help here
>> ( org.apache.juli.OneLineFormatter and %I in AccessLogValve)
>
> Probably not. Past experience is that it is the Processor that needs to
> be tracked. There should be a way to get that logged.
>
> As it happens some further digging showed that the failure did not
> depend on any other tests (I missed this the first time around as I was
> testing in a different environment where the test passed). Now I have a
> repeatable failure in a relatively simple test case I can start looking
> at it more closely. That is today's task.
If I can help around ajp let me know and send some details.
Regards,
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: Tomcat 7.0.33 progress report
Posted by Mark Thomas <ma...@apache.org>.
On 16/11/2012 04:03, Konstantin Kolinko wrote:
> 2012/11/16 Mark Thomas <ma...@apache.org>:
>> The 7.0.33 release is taking longer then planned as my pre-tag checks
>> are still ongoing.
>>
>> The short version is that the Servlet TCK is failing with httpd +
>> mod_proxy_ajp but not with httpd + mod_jk and I am still investigating.
>>
>> The long version is that I have taken a quick look at the failures but
>> it appears that the reported failures are symptoms rather than the root
>> cause (i.e. another test is getting Tomcat's internals into a bad state
>> and the next test that re-uses whichever object it is that is in a bad
>> state fails). Tracking down the test(s) triggering the problem will take
>> time. Since mod_jk is OK, it could be a mod_proxy_ajp bug we have hit
>> with the changes to the Async error handling. I have just updated the
>> httpd version I use for the TCKs and am re-running the TCK.
>>
>> I'll provide an updated progress report in a few days or just create the
>> tag if I resolve the issue.
>
> 1. I wonder whether configuring the logging to include a thread name
> will help here
> ( org.apache.juli.OneLineFormatter and %I in AccessLogValve)
Probably not. Past experience is that it is the Processor that needs to
be tracked. There should be a way to get that logged.
As it happens some further digging showed that the failure did not
depend on any other tests (I missed this the first time around as I was
testing in a different environment where the test passed). Now I have a
repeatable failure in a relatively simple test case I can start looking
at it more closely. That is today's task.
> 2. There is a bug in implementation of the "Out" tagplugin that should
> be fixed before tagging
> https://issues.apache.org/bugzilla/show_bug.cgi?id=54144#c2
>
> (I certainly do not have time for it for the next ~8 hours.)
Ack. I'm not going to get to that before the TCK issue is addressed. If
it still todo when I have finished the TCK stuff I'll take a look.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: Tomcat 7.0.33 progress report
Posted by Konstantin Kolinko <kn...@gmail.com>.
2012/11/16 Mark Thomas <ma...@apache.org>:
> The 7.0.33 release is taking longer then planned as my pre-tag checks
> are still ongoing.
>
> The short version is that the Servlet TCK is failing with httpd +
> mod_proxy_ajp but not with httpd + mod_jk and I am still investigating.
>
> The long version is that I have taken a quick look at the failures but
> it appears that the reported failures are symptoms rather than the root
> cause (i.e. another test is getting Tomcat's internals into a bad state
> and the next test that re-uses whichever object it is that is in a bad
> state fails). Tracking down the test(s) triggering the problem will take
> time. Since mod_jk is OK, it could be a mod_proxy_ajp bug we have hit
> with the changes to the Async error handling. I have just updated the
> httpd version I use for the TCKs and am re-running the TCK.
>
> I'll provide an updated progress report in a few days or just create the
> tag if I resolve the issue.
1. I wonder whether configuring the logging to include a thread name
will help here
( org.apache.juli.OneLineFormatter and %I in AccessLogValve)
2. There is a bug in implementation of the "Out" tagplugin that should
be fixed before tagging
https://issues.apache.org/bugzilla/show_bug.cgi?id=54144#c2
(I certainly do not have time for it for the next ~8 hours.)
Best regards,
Konstantin Kolinko
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org