After opening an issue, what should be expected next?
-----Original Message-----
From: Ryan Lubke [mailto:ryan.lubke_at_oracle.com]
Sent: Thursday, September 16, 2010 11:52 AM
To: users_at_glassfish.dev.java.net
Subject: Re: Interrupted idle thread
On 9/16/10 4:07 AM, Martin, Ray wrote:
> I have no listener changes. I did a search for " http
> request-timeout-seconds" to do as you had suggested earlier - but, did
> not find it in 3.0.1 domain.xml.
>
> If the test case is not repeatable by anyone else, then filing an
issue
> would be pointless - though I would be glad to do so if there is
value.
> Where does one go to open an issue?
I believe there to be value in logging an issue with the test case
(wouldn't have
asked otherwise :) ).
You can log an issue here:
https://glassfish.dev.java.net/servlets/ProjectIssues
> Thanx.
>
> -----Original Message-----
> From: Ryan Lubke [mailto:ryan.lubke_at_oracle.com]
> Sent: Wednesday, September 15, 2010 5:50 PM
> To: users_at_glassfish.dev.java.net
> Subject: Re: Interrupted idle thread
>
> On 9/15/10 1:16 PM, Martin, Ray wrote:
>> The test is very repeatable.
>>
>> I have two cases running right now - one on unpatched 3.0.1 and the
>> other on patched 3.0.1.
>>
>> Both cases show that a web service that runs over 15 minutes,
> glassfish
>> will wizz all over the parade.
>>
>> Both cases, right on 15 minutes every time, rip out and do very bad
>> things. Who ever heard of starting a second process because
something
>> went beyond 15 minutes. If one had a very sensitive process going -
>> like calculating and writing to a database - oop, yur database is all
>> screwed up - how nice.
>>
>> Patched 3.0.1 blows out a WebServiceException and starts a second
>> instance of the operation. This is the worst that could possibly
> occur.
>> Unpatched 3.0.1 only warns of "interrupted idle thread:...", starts
>> second instance of the operation, but quits the second instance
>> immediately. On an asynchronous operation, the client believes the
>> process has completed - this is certainly not a good thing, but at
> least
>> it did not corrupt my database.
> Hi Martin,
>
> At this point, I'd like to request you open an issue and attach your
> test case.
> Please include any relevant (or what you may think is relevant)
listener
>
> configuration
> changes you have made in the report.
>
> Thanks,
> -rl
>>
>> -----Original Message-----
>> From: Martin, Ray
>> Sent: Wednesday, September 15, 2010 2:54 PM
>> To: users_at_glassfish.dev.java.net
>> Subject: RE: RE: Re: Interrupted idle thread
>>
>> I ran the same test on a unpatched version of 3.0.1.
>> It also ran very badly - but, very differently. It kicked the
process
>> the second time, but immediately quits on the second process.
>> Also, it appears that the connection is still available between the
>> client and server - unlike previous version. Before, when getting a
>> ClientAbortException, the connection was no longer useful.
>>
>> I think I should go back to an unpatched version of 3.0.1 at a
> minimum.
>> -----Original Message-----
>> From: Martin, Ray
>> Sent: Wednesday, September 15, 2010 12:58 PM
>> To: users_at_glassfish.dev.java.net
>> Subject: RE: Re: Interrupted idle thread
>>
>> Now using glassfish 3.0.1
>>
>> Problem remains.
>>
>> Placed patch jars into glassfish directory.
>>
>> Still get the "interrupting idle thread...".
>>
>> Was always using a synchronous client up to this point.
>>
>> Created an asynchronous client - problem remains.
>>
>> I needed a more exact replication of the problem.
>>
>> I used Netbeans 6.9.1.
>>
>> Created a web service using session bean - the operations are setup
> and
>> activate.
>>
>> In the session bean, activate method, I put the following"
>>
>> System.out.println("Hello");
>> int min = 30;
>> long now = System.currentTimeMillis();
>> long wait = min*60*1000;
>> long go = now+wait;
>> min--;
>> while(now<go) {
>> try {
>> Thread.sleep(1000);
>> } catch ...
>> }
>> now = System.currentTimeMillis();
>> if(cnt>59) {
>> System.out.println(min--);
>> cnt = 0;
>> }
>> cnt++;
>> }
>>
>> The sys out is written to the glassfish log, of course. When the
>> Activate SOAP message arrives, Hello appears, one minute later a 29
>> appears. The sys out keeps counting down until a 16 appears.
>>
>> Then, WARNING: Interrupting idle Thread: http-thread-pool-8080-(2),
>> appears.
>>
>> Then Hello appears again, then a 15, then a 29, 14, 28
>>
>> Finally, the first minute count goes to 1 (the second count was 16)
>> Then the log shows:
>> WARNING: invocation error on ejb endpoint ...
>>
>> The second count continues to wind down to 1, then the error:
>> WARNING: invocation error on ejb endpoint ...
>>
>>
>> Note: I cannot find the " http request-timeout-seconds" in domain.xml
>>
>> Note: My brother says there is a mechanism to tell the client not to
>> timeout - he could not remember what it was. Anyone?
>>
>> Please help!
>>
>> Thanx,
>> Ray
>>
>>
>> -----Original Message-----
>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>> Sent: Thursday, September 02, 2010 12:54 PM
>> To: users_at_glassfish.dev.java.net
>> Subject: Re: Interrupted idle thread
>>
>> Martin,
>>
>> here is one more issue been fixed after 3.0.1 release, might be
>> related [1].
>> At the bottom you may find the proposed patch.
>>
>> WBR,
>> Alexey.
>>
>> [1] https://grizzly.dev.java.net/issues/show_bug.cgi?id=862
>>
>> On Sep 2, 2010, at 14:06 , Martin, Ray wrote:
>>
>>> Grizzly Framework 1.9.18-o
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>> Sent: Wednesday, September 01, 2010 5:08 PM
>>> To: users_at_glassfish.dev.java.net
>>> Subject: Re: Interrupted idle thread
>>>
>>> Can you pls. check the server.log file, there should be the line
like
>>> "Starting Grizzly Framework 1.9......."
>>>
>>> which exactly version you see there?
>>>
>>> Thanks.
>>> Alexey.
>>>
>>> On Sep 1, 2010, at 22:53 , Martin, Ray wrote:
>>>
>>>> Glassfish V3 also does another querky thing.
>>>>
>>>> When the process finishes running - in v2 it would send back the
> SOAP
>>>> response.
>>>> In v3, I get "WARNING: invocation error on ejb endpoint:
>>>> ConductorWebService at ConductorWebServiceSerice/
>>>> ConductorWebService :
>>>> com.ctc.wstx.exc.WstxIOException: null
>>>>
>>>> Ya gotta luv v2 - no such hassle :)
>>>>
>>>> -----Original Message-----
>>>> From: Martin, Ray
>>>> Sent: Wednesday, September 01, 2010 2:59 PM
>>>> To: users_at_glassfish.dev.java.net
>>>> Subject: RE: Re: Interrupted idle thread
>>>>
>>>> I would be glad to do so - I just have not figured out how just
yet.
>>>>
>>>> -----Original Message-----
>>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>>> Sent: Wednesday, September 01, 2010 2:57 PM
>>>> To: users_at_glassfish.dev.java.net
>>>> Subject: Re: Interrupted idle thread
>>>>
>>>> I see.
>>>> If you can provide some simple testcase to reproduce the issue -
> I'll
>>>> try to figure out what's going wrong.
>>>>
>>>> WBR,
>>>> Alexey.
>>>>
>>>> On Sep 1, 2010, at 20:30 , Martin, Ray wrote:
>>>>
>>>>> Yes, the netbeans auto generation code now uses an AbstractFacade
-
>>>>> it
>>>>> is gonna take some new understanding.
>>>>>
>>>>> V3.0.1 is still broken with respect to JAX-WS. I didn't measure
> the
>>>>> time just yet. But, when a process is activated via a JAX-WS
>>>>> command,
>>>>> V3 eventually does the "WARNING: Interrupting idle Thread:
>>>>> http-thread-pool-8080-(2)" - which does not mean a whole lot to
me.
>>>>>
>>>>> BUT, then V3 does a very bad thing - it acts as though it had
>>>>> received a
>>>>> second JAX-WS message and my process is activated the second time
-
>>>>> now
>>>>> I have two processes pouring junk into my database.
>>>>>
>>>>> I MUST go back to v2 - it worked - I have been running and testing
>>>>> these
>>>>> routines for months. Switching to V3 has crushed me.
>>>>>
>>>>> -----Original Message-----
>>>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>>>> Sent: Wednesday, September 01, 2010 4:47 AM
>>>>> To: users_at_glassfish.dev.java.net
>>>>> Subject: Re: Interrupted idle thread
>>>>>
>>>>> Hi Martin,
>>>>>
>>>>> didn't expect NB 6.9.1 makes something different than 6.8. Hope
>>>>> you'll
>>>>> finally make things work with it.
>>>>>
>>>>> Alexey.
>>>>>
>>>>> On Aug 31, 2010, at 18:29 , Martin, Ray wrote:
>>>>>
>>>>>> The javaee-api-6.0.jar answered the redlines in my prior code.
>>>>>>
>>>>>> I retested the code generators in netbeans - wow, life is
> different
>>>>>> than
>>>>>> the previous version - looks like I have lots of new learning to
>>>>>> do. I
>>>>>> guess this is all called progress.
>>>>>>
>>>>>> Tore the databases down in preparation for rebuilding.
>>>>>>
>>>>>> Deployed and ran - got an ORA-02049 - this looks like a problem
in
>>>>>> developer space!!! Ball is back in my court.
>>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Martin, Ray
>>>>>> Sent: Tuesday, August 31, 2010 12:19 PM
>>>>>> To: users_at_glassfish.dev.java.net
>>>>>> Subject: RE: RE: Re: Interrupted idle thread
>>>>>>
>>>>>> I figured those wonderful code generators in netbeans would tell
> me
>>>>>> what/how to do annotations. Wrong - the code generators did not
>>>>>> write
>>>>>> code like they use to. Don't know if it is because of the missing
>>>>>> libraries or if they just don't write code no more???
>>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Martin, Ray
>>>>>> Sent: Tuesday, August 31, 2010 12:06 PM
>>>>>> To: users_at_glassfish.dev.java.net
>>>>>> Subject: RE: Re: Interrupted idle thread
>>>>>>
>>>>>> I lucked out - we have netbeans 6.9.1
>>>>>>
>>>>>> I have installed.
>>>>>>
>>>>>> At admin console, I again setup all my connections pools and jdbc
>>>>>> resource. Did not yet do my security realm.
>>>>>>
>>>>>> Had to reset the server on my projects in netbeans.
>>>>>>
>>>>>> Everything is good so far - then oops - where are the EJB
>>>>>> libraries?
>>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>>>>> Sent: Tuesday, August 31, 2010 10:46 AM
>>>>>> To: users_at_glassfish.dev.java.net
>>>>>> Subject: Re: Interrupted idle thread
>>>>>>
>>>>>> It should work, but in general 3.0.1 is more stable, that's why
> IMO
>>>>>> it
>>>>>> could make sense to try it.
>>>>>>
>>>>>> Thanks.
>>>>>> Alexey.
>>>>>>
>>>>>> On Aug 31, 2010, at 16:21 , Martin, Ray wrote:
>>>>>>
>>>>>>> Okay - I will make the switch to a newer version ASAP.
>>>>>>>
>>>>>>> BUT - to get me back up and running, could I not set the other
>>>>>>> variable???
>>>>>>>
>>>>>>>
>>>>>>> The link that you provided says:
>>>>>>> Additional comments
>>>>>>> "We misinterpret thread-pool configuration elements in
> domain.xml,
>>>>>>> we
>>>>>>> use thread-pool->idle-thread-
>>>>>>> timeout-seconds to set transaction timeout and http->timeout-
>>>>>>> seconds
>>>>>>> to
>>>>>>> set thread's idle-timeout."
>>>>>>>
>>>>>>>
>>>>>>> So, if I knew the syntax, I could set
>>>>>>> thread-pool-idle-thread-timeout-seconds
>>>>>>> Just for enough to get me going - I am in a really bad position
> at
>>>>>>> the
>>>>>>> moment.
>>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>>>>>> Sent: Tuesday, August 31, 2010 9:58 AM
>>>>>>> To: users_at_glassfish.dev.java.net
>>>>>>> Subject: Re: Interrupted idle thread
>>>>>>>
>>>>>>> Now I see, most probably you're using GF v3, where we had issue
>>>>>>> [1].
>>>>>>> Can you pls. switch to GF v3.0.1?
>>>>>>>
>>>>>>> I don't remember exactly if NB 6.8 supports GF v3 server out of
>>>>>>> the
>>>>>>> box, or you need additional plugin to be installed, but with NB
>>>>>>> 6.9.1
>>>>>>> you can use GF 3.0.1 (AFAIR it's one you have out of the box),
>>>>>>> which
>>>>>>> should have that issue resolved.
>>>>>>>
>>>>>>> Thanks.
>>>>>>>
>>>>>>> WBR,
>>>>>>> Alexey.
>>>>>>>
>>>>>>> [1] https://glassfish.dev.java.net/issues/show_bug.cgi?id=11490
>>>>>>>
>>>>>>>
>>>>>>> On Aug 31, 2010, at 14:21 , Martin, Ray wrote:
>>>>>>>
>>>>>>>> I am using netbeans6.8 and whatever glassfish comes with it.
>>>>>>>>
>>>>>>>> At the glassfish admin console, I press the version button, I
> get
>>>>>>>> v3
>>>>>>>> (no
>>>>>>>> numbers)
>>>>>>>>
>>>>>>>> I am very sorry to say, that I cannot get the descriptor onto
> the
>>>>>>>> Internet.
>>>>>>>>
>>>>>>>> -----Original Message-----
>>>>>>>> From: Oleksiy.Stashok_at_Sun.COM [mailto:Oleksiy.Stashok_at_Sun.COM]
>>>>>>>> Sent: Tuesday, August 31, 2010 7:35 AM
>>>>>>>> To: users_at_glassfish.dev.java.net
>>>>>>>> Subject: Re: Interrupted idle thread
>>>>>>>>
>>>>>>>> Martin,
>>>>>>>>
>>>>>>>> just to make sure, are you using the Glassfish 3.0.1 release?
>>>>>>>> Can you pls. share the domain.xml descriptor?
>>>>>>>>
>>>>>>>> Thanks.
>>>>>>>>
>>>>>>>> Alexey.
>>>>>>>>
>>>>>>>> On Aug 31, 2010, at 13:28 , Martin, Ray wrote:
>>>>>>>>
>>>>>>>>> Mr. Lubke,
>>>>>>>>>
>>>>>>>>> I did the following:
>>>>>>>>> ...
>>>>>>>>> <http request-timeout-seconds="-1" ...
>>>>>>>>> ...
>>>>>>>>>
>>>>>>>>> I assume that -1 means to disable.
>>>>>>>>>
>>>>>>>>> After restarting glassfish and deploying the EJBModule in an
>>>>>>>>> Enterprise
>>>>>>>>> app (so it will see the third party jars) - still got the
>>>>>>>>> There are no highlighted small blocks.>>>>>>>
>>>>>>>>> But, worse than that is the fact that something causes the
SOAP
>>>>>>>>> message
>>>>>>>>> to be repeated - then my process is not protected against that
>>>>>>>>> repeat
>>>>>>>>> SOAP message again kicks off and stomps all over the database
>>>>>>>>> records
>>>>>>>>> that it just wrote before the timeout.
>>>>>>>>>
>>>>>>>>> I am going to try an asynchronous client to see if that will
>>>>>>>>> take
>>>>>>>>> care
>>>>>>>>> of the problem.
>>>>>>>>>
>>>>>>>>> Thanx for your help.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> -----Original Message-----
>>>>>>>>> From: Ryan Lubke [mailto:ryan.lubke_at_oracle.com]
>>>>>>>>> Sent: Monday, August 30, 2010 8:51 PM
>>>>>>>>> To: Martin, Ray
>>>>>>>>> Cc: users_at_glassfish.dev.java.net
>>>>>>>>> Subject: RE: Interrupted idle thread
>>>>>>>>>
>>>>>>>>> Hello Martin,
>>>>>>>>>
>>>>>>>>> I believe you'll be able to resolve this issue by either
>>>>>>>>> increasing
>>>>>>>>> the
>>>>>>>>> request-timeout-seconds property or disabling
>>>>>>>>> the feature altogether.
>>>>>>>>>
>>>>>>>>> This can be done by backing up your
>>>>>>>>> GF_HOME/domains/domain1/config/domain.xml and then modify the
>>>>>>>>> http
>>>>>>>>> element
>>>>>>>>> associated with the protocol your listener is using.
>>>>>>>>>
>>>>>>>>> For example, if you're using the out-of-the-box configuration,
>>>>>>>>> you
>>>>>>>>> can
>>>>>>>>> change the timeout for requests made to port 8080 like
>>>>>>>>> so:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> <protocol name="http-listener-1">
>>>>>>>>> <http request-timeout-seconds="60" default-virtual-
>>>>>>>>> server="server"
>>>>>>>>> max-connections="250" />
>>>>>>>>> </protocol>
>>>>>>>>>
>>>>>>>>> It doesn't appear that this value can currently be set to -1
> via
>>>>>>>>> asadmin
>>>>>>>>>
>>>>>>>>> command line, so I'll be looking into resolving
>>>>>>>>> that shortly.
>>>>>>>>>
>>>>>>>>> Could you please try updating the domain.xml using the above
> xml
>>>>>>>>> snippet
>>>>>>>>>
>>>>>>>>> as a guide and let us know if it resolves your problem?
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> -rl
>>>>>>>>>
>>>>>>>>>
>>>
---------------------------------------------------------------------
>>>>>>>>> To unsubscribe, e-mail:
> users-unsubscribe_at_glassfish.dev.java.net
>>>>>>>>> For additional commands, e-mail: users-
>>>>>>>>> help_at_glassfish.dev.java.net
>>>>>>>>>
>>>>>>>>
>> ---------------------------------------------------------------------
>>>>>>>> To unsubscribe, e-mail:
users-unsubscribe_at_glassfish.dev.java.net
>>>>>>>> For additional commands, e-mail: users-
>>>>>>>> help_at_glassfish.dev.java.net
>>>>>>>>
>>>>>>>>
>>>>>>>>
>> ---------------------------------------------------------------------
>>>>>>>> To unsubscribe, e-mail:
users-unsubscribe_at_glassfish.dev.java.net
>>>>>>>> For additional commands, e-mail: users-
>>>>>>>> help_at_glassfish.dev.java.net
>>>>>>>>
>>>>>>>
> ---------------------------------------------------------------------
>>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>>> For additional commands, e-mail:
> users-help_at_glassfish.dev.java.net
>>>>>>>
>>>>>>>
> ---------------------------------------------------------------------
>>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>>> For additional commands, e-mail:
> users-help_at_glassfish.dev.java.net
>>>>>>
>>>
---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>> For additional commands, e-mail:
users-help_at_glassfish.dev.java.net
>>>>>>
>>>>>>
>>>>>>
>>>
---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>> For additional commands, e-mail:
users-help_at_glassfish.dev.java.net
>>>>>>
>>>>>>
>>>>>>
>>>
---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>> For additional commands, e-mail:
users-help_at_glassfish.dev.java.net
>>>>>>
>>>>>>
>>>>>>
>>>
---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>>> For additional commands, e-mail:
users-help_at_glassfish.dev.java.net
>>>>>>
>>>>>
>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>>
>>>>>
>>>>>
>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>>
>>>>
> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>
>>>>
>>>>
> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>
>>>>
>>>>
> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>
>>>
---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>
>>>
>>>
---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
For additional commands, e-mail: users-help_at_glassfish.dev.java.net