Re: [OPENEJB-1303] JMX API Tests

["viola.lu" <vi...@gmail.com>]

I already provided a patch for
https://issues.apache.org/jira/browse/OPENEJB-1303, seems feed won't catch
thi update, can you help review it?
And there is a problem here, may bug or not:
Set replaced aged as true via:
statelessContainerInfo.properties.setProperty("ReplaceAged", "true"); , but
its attribute value is still false by getting mbeanserver attribute
value://ReplaceAged false, not sure whether it's openejb issue or its
attribute name is right? 
Thanks in advance!
-- 
View this message in context: http://openejb.979440.n4.nabble.com/OPENEJB-1303-JMX-API-Tests-tp2266444p2278972.html
Sent from the OpenEJB Dev mailing list archive at Nabble.com.

Re: [OPENEJB-1303] JMX API Tests

["viola.lu" <vi...@gmail.com>]

Hi, david:
 Thanks.I will update geronimo StatelessContainerGBean to reflect those
properties name change. As you know, openejb 3.1.3-SNAPSHOT is integrated
into geornimo 2.2.2-snaphot, can you merge those code update in openejb
trunk into openejb 3.1.3 branch? So geronimo can utilize latests update.
Thanks in advance!

-- 
View this message in context: http://openejb.979440.n4.nabble.com/OPENEJB-1303-JMX-API-Tests-tp2266444p2283035.html
Sent from the OpenEJB Dev mailing list archive at Nabble.com.

Re: [OPENEJB-1303] JMX API Tests

[David Blevins <da...@visi.com>]

On Jul 8, 2010, at 3:18 AM, viola.lu wrote:

> For accessTimeouts test, there will be an accesstimeout exception if
> invocation against a full pool:
> javax.ejb.ConcurrentAccessTimeoutException: No instances available in
> Stateless Session Bean pool.  Waited 1 MILLISECONDS.
> For idletimouts test, i tried serveral way to check it, but it always keep
> 0L, not sure why,from the comments: 
> Fill a pool to the max, let the non-min instances timeout, check the
> IdleTimeouts, but after fill a max pull, non-min instances timeout
> For aging, replaceAged setting still keep false.
> 
> i attached the patch on jira
> https://issues.apache.org/jira/browse/OPENEJB-1303
> pls help review it,

Thanks, Viola!  I had some related changes already from fixing stats bugs, so I merged that patch in with my last commit with some modifications.

I really like the actualDate.after(expected) technique in the sweeps test.  That was super clever.  It's so simple and effective I just went ahead and removed the year checking part since the "after" part is doing that and more.  Going to have to remember that if I ever need to do some date testing.

> thanks your help of improve my code in previous thread,
> which make me learn more how to keep code simple and easy to read.

No problem!  It's been fun working on the tests with you.  I'm really proud of our results.  They look great and I never would have had time to do all alone.

We're good on the stats and bugs now, but there's still a little work left on the Geronimo side.  I added and renamed a few config properties, so we'll need to update the related Geronimo gbeans/plans to reflect.  If you have time to help with that, it'd be wonderful.

Thanks for all the great work, Viola!

-David

Re: [OPENEJB-1303] JMX API Tests

["viola.lu" <vi...@gmail.com>]

Hi, david:
  For accessTimeouts test, there will be an accesstimeout exception if
invocation against a full pool:
javax.ejb.ConcurrentAccessTimeoutException: No instances available in
Stateless Session Bean pool.  Waited 1 MILLISECONDS.
 For idletimouts test, i tried serveral way to check it, but it always keep
0L, not sure why,from the comments: 
 Fill a pool to the max, let the non-min instances timeout, check the
IdleTimeouts, but after fill a max pull, non-min instances timeout
For aging, replaceAged setting still keep false.

i attached the patch on jira
https://issues.apache.org/jira/browse/OPENEJB-1303
pls help review it, thanks your help of improve my code in previous thread,
which make me learn more how to keep code simple and easy to read.

 


-- 
View this message in context: http://openejb.979440.n4.nabble.com/OPENEJB-1303-JMX-API-Tests-tp2266444p2282087.html
Sent from the OpenEJB Dev mailing list archive at Nabble.com.

Re: [OPENEJB-1303] JMX API Tests

[David Blevins <da...@visi.com>]

Hi Viola,

On Jul 5, 2010, at 6:00 PM, viola.lu wrote:

> I already provided a patch for
> https://issues.apache.org/jira/browse/OPENEJB-1303, seems feed won't catch
> thi update, can you help review it?

Looks like a great start!  Went ahead and applied it.  Thank you!

> And there is a problem here, may bug or not:
> Set replaced aged as true via:
> statelessContainerInfo.properties.setProperty("ReplaceAged", "true"); , but
> its attribute value is still false by getting mbeanserver attribute
> value://ReplaceAged false, not sure whether it's openejb issue or its
> attribute name is right? 

Sounds like a bug.  As you note in the JIRA PollInterval also seems wrong.  We definitely should have a test that changes all the pool configuration settings on the container and verifies that those settings are correct in the stats.  

There seems to also be a bug with the "Latest" times as well.  We should pick one of the "Latest" stats and test it a bit.  We should at least reparse the date and at least check the day, month and year are correct.  Currently it says 1970 or something strange as the year which definitely isn't right.

I split the pool related stats tests into a different test class and moved the javadoc at the top down onto test methods which still need to be implemented.  I fixed PoolVersion so that it reports as an int rather than AtomicInteger -- that was also a bug.

Thanks for all the work on this!  It's clear we have some bugs to flush out before calling this code finished.  The help is greatly appreciated!


-David