4:55 p.m.
Hi,
I found a cause of memory leaks in our test suite [1]. I corrected easy
fixes but I did not go through individual test classes that need to be
updated (see attachment in JIRA). I thought it was best to leave test
updating to devs that are most familiar with particular test classes.
Best regards,
Vladimir
[1] https://jira.jboss.org/jira/browse/ISPN-220
2:34 a.m.
Hi Vladimir
Thanks for looking into this. Can't we add some hooks into
UnitTestTestNGListener to clean up the test class, setting internal
refs to null using reflection? This would then be pretty non-
intrusive to the actual test...
Cheers
Manik
On 7 Oct 2009, at 22:55, Vladimir Blagojevic wrote:
Hi,
I found a cause of memory leaks in our test suite [1]. I corrected
easy
fixes but I did not go through individual test classes that need to be
updated (see attachment in JIRA). I thought it was best to leave test
updating to devs that are most familiar with particular test classes.
Best regards,
Vladimir
[1] https://jira.jboss.org/jira/browse/ISPN-220
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
10:16 a.m.
Or we can put it in AbstractCacheTest. Great idea either way!
On 09-10-09 3:34 AM, Manik Surtani wrote:
Hi Vladimir
Thanks for looking into this. Can't we add some hooks into
UnitTestTestNGListener to clean up the test class, setting internal
refs to null using reflection? This would then be pretty non-
intrusive to the actual test...
Cheers
Manik
10:35 a.m.
Yes, but you said that superclass @AfterClass methods won't be invoked
anyway ... so UnitTestTestNGListener probably makes more sense.
On 9 Oct 2009, at 16:16, Vladimir Blagojevic wrote:
Or we can put it in AbstractCacheTest. Great idea either way!
On 09-10-09 3:34 AM, Manik Surtani wrote:
> Hi Vladimir
>
> Thanks for looking into this. Can't we add some hooks into
> UnitTestTestNGListener to clean up the test class, setting internal
> refs to null using reflection? This would then be pretty non-
> intrusive to the actual test...
>
> Cheers
> Manik
>
>
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
10:39 a.m.
On 09-10-09 11:35 AM, Manik Surtani wrote:
Yes, but you said that superclass @AfterClass methods won't be
invoked
anyway ... so UnitTestTestNGListener probably makes more sense.
TestNG starts from the leaf class and searches for a first class that
has a method annotated with @AfterClass and invokes it. It does not
continue to climb the hierarchy. However, all needs to be done is to
invoke super from SingleCacheManagerTest and MultipleCacheManagerTest.
However, that said, we are still left with a bunch of test classes that
do not inherit from AbstractCacheTest :(
So yes, I'll see how to hookup UnitTestTestNGListener :)
Cheers
On 9 Oct 2009, at 16:16, Vladimir Blagojevic wrote:
> Or we can put it in AbstractCacheTest. Great idea either way!
>
> On 09-10-09 3:34 AM, Manik Surtani wrote:
>
>> Hi Vladimir
>>
>> Thanks for looking into this. Can't we add some hooks into
>> UnitTestTestNGListener to clean up the test class, setting internal
>> refs to null using reflection? This would then be pretty non-
>> intrusive to the actual test...
>>
>> Cheers
>> Manik
>>
>>
>>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
3:36 p.m.
On 09-10-09 11:39 AM, Vladimir Blagojevic wrote:
TestNG starts from the leaf class and searches for a first class
that
has a method annotated with @AfterClass and invokes it. It does not
continue to climb the hierarchy. However, all needs to be done is to
invoke super from SingleCacheManagerTest and MultipleCacheManagerTest.
However, that said, we are still left with a bunch of test classes that
do not inherit from AbstractCacheTest :(
So yes, I'll see how to hookup UnitTestTestNGListener :)
Hard to believe but there is no way to do this through listener. After I
could not find a way to do it I even posted a question on testng-users
and got a recommendation from Cedric to nullify all instances from a
@AfterClass method in superclass.
I also found that I was wrong regarding invocation of @AfterClass
methods:) As long as you do not override @AfterClass method it *will* be
invoked starting from the leaf all the way up through ancestors. I
created a superclass called AbstractInfinispanTest with
nullifyInstanceFields method and made AbstractInfinispanTest superclass
of all test classes. Works as a charm and guranteed no more leaks!
Let me know if you have any reservations to this change.
Vladimir
4:53 a.m.
On 9 Oct 2009, at 21:36, Vladimir Blagojevic wrote:
On 09-10-09 11:39 AM, Vladimir Blagojevic wrote:
> TestNG starts from the leaf class and searches for a first class that
> has a method annotated with @AfterClass and invokes it. It does not
> continue to climb the hierarchy. However, all needs to be done is to
> invoke super from SingleCacheManagerTest and
> MultipleCacheManagerTest.
>
> However, that said, we are still left with a bunch of test classes
> that
> do not inherit from AbstractCacheTest :(
>
> So yes, I'll see how to hookup UnitTestTestNGListener :)
>
Hard to believe but there is no way to do this through listener.
After I
could not find a way to do it I even posted a question on testng-users
and got a recommendation from Cedric to nullify all instances from a
@AfterClass method in superclass.
I also found that I was wrong regarding invocation of @AfterClass
methods:) As long as you do not override @AfterClass method it
*will* be
invoked starting from the leaf all the way up through ancestors. I
created a superclass called AbstractInfinispanTest with
nullifyInstanceFields method and made AbstractInfinispanTest
superclass
of all test classes. Works as a charm and guranteed no more leaks!
Let me know if you have any reservations to this change.
As long as it works and is non-intrusive, I'm happy. :-)
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
12:20 a.m.
On 09-10-11 5:53 AM, Manik Surtani wrote:
As long as it works and is non-intrusive, I'm happy. :-)
Done. There has been a regression in our test suite that I believe is
unrelated to the change I made regarding these leaks. I did not pick up
commits 936 and 937 and I do not see new failures [1]. So you guys have
a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
worked on pass if test suite is invoked from core directory but the fail
if invoked from infinispan directory - as hudson does.
[1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/812/
http://fisheye.jboss.org/changelog/Infinispan/?cs=937
http://fisheye.jboss.org/changelog/Infinispan/?cs=936
4:30 a.m.
Great, thanks for looking into this!
On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
On 09-10-11 5:53 AM, Manik Surtani wrote:
> As long as it works and is non-intrusive, I'm happy. :-)
Done. There has been a regression in our test suite that I believe is
unrelated to the change I made regarding these leaks. I did not pick
up
commits 936 and 937 and I do not see new failures [1]. So you guys
have
a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
worked on pass if test suite is invoked from core directory but the
fail
if invoked from infinispan directory - as hudson does.
[1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/812/
http://fisheye.jboss.org/changelog/Infinispan/?cs=937
http://fisheye.jboss.org/changelog/Infinispan/?cs=936
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
11:22 a.m.
Absolutely, well done Vladimir :)
On 10/12/2009 11:30 AM, Manik Surtani wrote:
Great, thanks for looking into this!
On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
> On 09-10-11 5:53 AM, Manik Surtani wrote:
>> As long as it works and is non-intrusive, I'm happy. :-)
> Done. There has been a regression in our test suite that I believe is
> unrelated to the change I made regarding these leaks. I did not pick
> up
> commits 936 and 937 and I do not see new failures [1]. So you guys
> have
> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
> worked on pass if test suite is invoked from core directory but the
> fail
> if invoked from infinispan directory - as hudson does.
>
>
> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/812/
>
>
> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
10:12 p.m.
UDP build is also up and running, no more problems for members to find
each other. Hudson did not like binding to 127.0.0.1 for mcast sockets....
On 09-10-12 5:30 AM, Manik Surtani wrote:
Great, thanks for looking into this!
On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
> On 09-10-11 5:53 AM, Manik Surtani wrote:
>
>> As long as it works and is non-intrusive, I'm happy. :-)
>>
> Done. There has been a regression in our test suite that I believe is
> unrelated to the change I made regarding these leaks. I did not pick
> up
> commits 936 and 937 and I do not see new failures [1]. So you guys
> have
> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
> worked on pass if test suite is invoked from core directory but the
> fail
> if invoked from infinispan directory - as hudson does.
>
>
> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/812/
>
>
> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
4:45 a.m.
On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
UDP build is also up and running, no more problems for members to
find
each other. Hudson did not like binding to 127.0.0.1 for mcast
sockets....
You're on a roll! So what are you actually binding to now?
BTW we seem to have a stuck test when using UDP...
http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
On 09-10-12 5:30 AM, Manik Surtani wrote:
> Great, thanks for looking into this!
>
> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>
>
>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>
>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>
>> Done. There has been a regression in our test suite that I believe
>> is
>> unrelated to the change I made regarding these leaks. I did not pick
>> up
>> commits 936 and 937 and I do not see new failures [1]. So you guys
>> have
>> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
>> worked on pass if test suite is invoked from core directory but the
>> fail
>> if invoked from infinispan directory - as hudson does.
>>
>>
>> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>> 812/
>>
>>
>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
> --
> Manik Surtani
> manik(a)jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
4:54 a.m.
Thank you very much for sorting that out! I was really out of ideas on
what to do next to make it work.
Kudos to you!
On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
> UDP build is also up and running, no more problems for members to
> find
> each other. Hudson did not like binding to 127.0.0.1 for mcast
> sockets....
You're on a roll! So what are you actually binding to now?
BTW we seem to have a stuck test when using UDP...
http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>
> On 09-10-12 5:30 AM, Manik Surtani wrote:
>> Great, thanks for looking into this!
>>
>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>
>>
>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>
>>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>>
>>> Done. There has been a regression in our test suite that I believe
>>> is
>>> unrelated to the change I made regarding these leaks. I did not
>>> pick
>>> up
>>> commits 936 and 937 and I do not see new failures [1]. So you guys
>>> have
>>> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
>>> worked on pass if test suite is invoked from core directory but the
>>> fail
>>> if invoked from infinispan directory - as hudson does.
>>>
>>>
>>> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>> 812/
>>>
>>>
>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>> --
>> Manik Surtani
>> manik(a)jboss.org
>> Lead, Infinispan
>> Lead, JBoss Cache
>> http://www.infinispan.org
>> http://www.jbosscache.org
>>
>>
>>
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
5:05 a.m.
It must be something related to either:
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/resources/...
or
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/java/org/i...
bind.address was the old pre-JGroups 2.4.x format and jgroups.bind_addr
the new one since JGroups 2.4.x.
On 10/13/2009 11:54 AM, Mircea Markus wrote:
Thank you very much for sorting that out! I was really out of ideas
on
what to do next to make it work.
Kudos to you!
On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
>
> On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
>
>> UDP build is also up and running, no more problems for members to
>> find
>> each other. Hudson did not like binding to 127.0.0.1 for mcast
>> sockets....
>
> You're on a roll! So what are you actually binding to now?
>
> BTW we seem to have a stuck test when using UDP...
>
> http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>
>>
>> On 09-10-12 5:30 AM, Manik Surtani wrote:
>>> Great, thanks for looking into this!
>>>
>>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>>
>>>
>>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>>
>>>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>>>
>>>> Done. There has been a regression in our test suite that I believe
>>>> is
>>>> unrelated to the change I made regarding these leaks. I did not
>>>> pick
>>>> up
>>>> commits 936 and 937 and I do not see new failures [1]. So you guys
>>>> have
>>>> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
>>>> worked on pass if test suite is invoked from core directory but the
>>>> fail
>>>> if invoked from infinispan directory - as hudson does.
>>>>
>>>>
>>>> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>>> 812/
>>>>
>>>>
>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>> --
>>> Manik Surtani
>>> manik(a)jboss.org
>>> Lead, Infinispan
>>> Lead, JBoss Cache
>>> http://www.infinispan.org
>>> http://www.jbosscache.org
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Manik Surtani
> manik(a)jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
5:09 a.m.
On 13 Oct 2009, at 11:05, Galder Zamarreno wrote:
It must be something related to either:
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/resources/...
or
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/java/org/i...
bind.address was the old pre-JGroups 2.4.x format and
jgroups.bind_addr
the new one since JGroups 2.4.x.
Any volunteers to fix this?
On 10/13/2009 11:54 AM, Mircea Markus wrote:
> Thank you very much for sorting that out! I was really out of ideas
> on
> what to do next to make it work.
> Kudos to you!
>
> On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
>
>>
>> On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
>>
>>> UDP build is also up and running, no more problems for members to
>>> find
>>> each other. Hudson did not like binding to 127.0.0.1 for mcast
>>> sockets....
>>
>> You're on a roll! So what are you actually binding to now?
>>
>> BTW we seem to have a stuck test when using UDP...
>>
>> http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>>
>>>
>>> On 09-10-12 5:30 AM, Manik Surtani wrote:
>>>> Great, thanks for looking into this!
>>>>
>>>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>>>
>>>>
>>>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>>>
>>>>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>>>>
>>>>> Done. There has been a regression in our test suite that I
>>>>> believe
>>>>> is
>>>>> unrelated to the change I made regarding these leaks. I did not
>>>>> pick
>>>>> up
>>>>> commits 936 and 937 and I do not see new failures [1]. So you
>>>>> guys
>>>>> have
>>>>> a look there. EhCache and jbc2InfinispanTransformerTest that
>>>>> Mircea
>>>>> worked on pass if test suite is invoked from core directory but
>>>>> the
>>>>> fail
>>>>> if invoked from infinispan directory - as hudson does.
>>>>>
>>>>>
>>>>> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>>>> 812/
>>>>>
>>>>>
>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>
>>>> --
>>>> Manik Surtani
>>>> manik(a)jboss.org
>>>> Lead, Infinispan
>>>> Lead, JBoss Cache
>>>> http://www.infinispan.org
>>>> http://www.jbosscache.org
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> --
>> Manik Surtani
>> manik(a)jboss.org
>> Lead, Infinispan
>> Lead, JBoss Cache
>> http://www.infinispan.org
>> http://www.jbosscache.org
>>
>>
>>
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
5:11 a.m.
On 10/13/2009 12:09 PM, Manik Surtani wrote:
On 13 Oct 2009, at 11:05, Galder Zamarreno wrote:
> It must be something related to either:
>
>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/resources/...
>
> or
>
>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/java/org/i...
>
> bind.address was the old pre-JGroups 2.4.x format and
> jgroups.bind_addr
> the new one since JGroups 2.4.x.
Any volunteers to fix this?
Fix what? Are you assuming that the old bind.address is not working any
more? AFAIK, it has so far been working but maybe something inside of
JGroups only works with the new system property? Or Vladimir/Bela, old
system properties not supported any more? I haven't seen any emails on that.
>
> On 10/13/2009 11:54 AM, Mircea Markus wrote:
>> Thank you very much for sorting that out! I was really out of ideas
>> on
>> what to do next to make it work.
>> Kudos to you!
>>
>> On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
>>
>>>
>>> On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
>>>
>>>> UDP build is also up and running, no more problems for members to
>>>> find
>>>> each other. Hudson did not like binding to 127.0.0.1 for mcast
>>>> sockets....
>>>
>>> You're on a roll! So what are you actually binding to now?
>>>
>>> BTW we seem to have a stuck test when using UDP...
>>>
>>>
http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>>>
>>>>
>>>> On 09-10-12 5:30 AM, Manik Surtani wrote:
>>>>> Great, thanks for looking into this!
>>>>>
>>>>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>>>>
>>>>>
>>>>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>>>>
>>>>>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>>>>>
>>>>>> Done. There has been a regression in our test suite that I
>>>>>> believe
>>>>>> is
>>>>>> unrelated to the change I made regarding these leaks. I did not
>>>>>> pick
>>>>>> up
>>>>>> commits 936 and 937 and I do not see new failures [1]. So you
>>>>>> guys
>>>>>> have
>>>>>> a look there. EhCache and jbc2InfinispanTransformerTest that
>>>>>> Mircea
>>>>>> worked on pass if test suite is invoked from core directory but
>>>>>> the
>>>>>> fail
>>>>>> if invoked from infinispan directory - as hudson does.
>>>>>>
>>>>>>
>>>>>> [1]
http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>>>>> 812/
>>>>>>
>>>>>>
>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>>>>> _______________________________________________
>>>>>> infinispan-dev mailing list
>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>
>>>>> --
>>>>> Manik Surtani
>>>>> manik(a)jboss.org
>>>>> Lead, Infinispan
>>>>> Lead, JBoss Cache
>>>>> http://www.infinispan.org
>>>>> http://www.jbosscache.org
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>> --
>>> Manik Surtani
>>> manik(a)jboss.org
>>> Lead, Infinispan
>>> Lead, JBoss Cache
>>> http://www.infinispan.org
>>> http://www.jbosscache.org
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Galder Zamarreño
> Sr. Software Engineer
> Infinispan, JBoss Cache
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
5:16 a.m.
On 13 Oct 2009, at 11:11, Galder Zamarreno wrote:
On 10/13/2009 12:09 PM, Manik Surtani wrote:
>
> On 13 Oct 2009, at 11:05, Galder Zamarreno wrote:
>
>> It must be something related to either:
>>
>>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/resources/...
>>
>> or
>>
>>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/java/org/i...
>>
>> bind.address was the old pre-JGroups 2.4.x format and
>> jgroups.bind_addr
>> the new one since JGroups 2.4.x.
>
> Any volunteers to fix this?
Fix what? Are you assuming that the old bind.address is not working
any
more?
Yes, or will not work as JGroups progresses so it is worthwhile using
the new property.
AFAIK, it has so far been working but maybe something inside of
JGroups only works with the new system property? Or Vladimir/Bela, old
system properties not supported any more? I haven't seen any emails
on that.
>
>>
>> On 10/13/2009 11:54 AM, Mircea Markus wrote:
>>> Thank you very much for sorting that out! I was really out of ideas
>>> on
>>> what to do next to make it work.
>>> Kudos to you!
>>>
>>> On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
>>>
>>>>
>>>> On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
>>>>
>>>>> UDP build is also up and running, no more problems for members to
>>>>> find
>>>>> each other. Hudson did not like binding to 127.0.0.1 for mcast
>>>>> sockets....
>>>>
>>>> You're on a roll! So what are you actually binding to now?
>>>>
>>>> BTW we seem to have a stuck test when using UDP...
>>>>
>>>>
http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>>>>
>>>>>
>>>>> On 09-10-12 5:30 AM, Manik Surtani wrote:
>>>>>> Great, thanks for looking into this!
>>>>>>
>>>>>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>>>>>
>>>>>>
>>>>>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>>>>>
>>>>>>>> As long as it works and is non-intrusive, I'm happy.
:-)
>>>>>>>>
>>>>>>> Done. There has been a regression in our test suite that I
>>>>>>> believe
>>>>>>> is
>>>>>>> unrelated to the change I made regarding these leaks. I did
not
>>>>>>> pick
>>>>>>> up
>>>>>>> commits 936 and 937 and I do not see new failures [1]. So
you
>>>>>>> guys
>>>>>>> have
>>>>>>> a look there. EhCache and jbc2InfinispanTransformerTest that
>>>>>>> Mircea
>>>>>>> worked on pass if test suite is invoked from core directory
but
>>>>>>> the
>>>>>>> fail
>>>>>>> if invoked from infinispan directory - as hudson does.
>>>>>>>
>>>>>>>
>>>>>>> [1]
http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>>>>>> 812/
>>>>>>>
>>>>>>>
>>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>>>>>> _______________________________________________
>>>>>>> infinispan-dev mailing list
>>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>>
>>>>>> --
>>>>>> Manik Surtani
>>>>>> manik(a)jboss.org
>>>>>> Lead, Infinispan
>>>>>> Lead, JBoss Cache
>>>>>> http://www.infinispan.org
>>>>>> http://www.jbosscache.org
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> infinispan-dev mailing list
>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>>> --
>>>> Manik Surtani
>>>> manik(a)jboss.org
>>>> Lead, Infinispan
>>>> Lead, JBoss Cache
>>>> http://www.infinispan.org
>>>> http://www.jbosscache.org
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> --
>> Galder Zamarreño
>> Sr. Software Engineer
>> Infinispan, JBoss Cache
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Manik Surtani
> manik(a)jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
5:18 a.m.
Galder Zamarreno wrote:
Fix what? Are you assuming that the old bind.address is not working
any
more? AFAIK, it has so far been working but maybe something inside of
JGroups only works with the new system property? Or Vladimir/Bela, old
system properties not supported any more? I haven't seen any emails on
that.
Both bind.address and jgroups.bind_addr are still supported, but I
recommend switching to using the new property (jgroups.bind_addr)
--
Bela Ban
Lead JGroups / Clustering Team
JBoss
5:40 a.m.
On Oct 13, 2009, at 1:11 PM, Galder Zamarreno wrote:
On 10/13/2009 12:09 PM, Manik Surtani wrote:
>
> On 13 Oct 2009, at 11:05, Galder Zamarreno wrote:
>
>> It must be something related to either:
>>
>>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/resources/...
>>
>> or
>>
>>
http://fisheye.jboss.org/browse/Infinispan/trunk/core/src/test/java/org/i...
>>
>> bind.address was the old pre-JGroups 2.4.x format and
>> jgroups.bind_addr
>> the new one since JGroups 2.4.x.
>
> Any volunteers to fix this?
Fix what? Are you assuming that the old bind.address is not working
any
more? AFAIK, it has so far been working but maybe something inside of
JGroups only works with the new system property? Or Vladimir/Bela, old
system properties not supported any more? I haven't seen any emails
on that.
if that would be the case the tests would fail all the time. They used
to pass on the same box when being run from the console and fail only
when trigger by hudson process
>
>>
>> On 10/13/2009 11:54 AM, Mircea Markus wrote:
>>> Thank you very much for sorting that out! I was really out of ideas
>>> on
>>> what to do next to make it work.
>>> Kudos to you!
>>>
>>> On Oct 13, 2009, at 12:45 PM, Manik Surtani wrote:
>>>
>>>>
>>>> On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
>>>>
>>>>> UDP build is also up and running, no more problems for members to
>>>>> find
>>>>> each other. Hudson did not like binding to 127.0.0.1 for mcast
>>>>> sockets....
>>>>
>>>> You're on a roll! So what are you actually binding to now?
>>>>
>>>> BTW we seem to have a stuck test when using UDP...
>>>>
>>>>
http://hudson.qa.jboss.com/hudson/view/Infinispan/job/Infinispan-trunk-JD...
>>>>
>>>>>
>>>>> On 09-10-12 5:30 AM, Manik Surtani wrote:
>>>>>> Great, thanks for looking into this!
>>>>>>
>>>>>> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>>>>>>
>>>>>>
>>>>>>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>>>>>>
>>>>>>>> As long as it works and is non-intrusive, I'm happy.
:-)
>>>>>>>>
>>>>>>> Done. There has been a regression in our test suite that I
>>>>>>> believe
>>>>>>> is
>>>>>>> unrelated to the change I made regarding these leaks. I did
not
>>>>>>> pick
>>>>>>> up
>>>>>>> commits 936 and 937 and I do not see new failures [1]. So
you
>>>>>>> guys
>>>>>>> have
>>>>>>> a look there. EhCache and jbc2InfinispanTransformerTest that
>>>>>>> Mircea
>>>>>>> worked on pass if test suite is invoked from core directory
but
>>>>>>> the
>>>>>>> fail
>>>>>>> if invoked from infinispan directory - as hudson does.
>>>>>>>
>>>>>>>
>>>>>>> [1]
http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>>>>>>> 812/
>>>>>>>
>>>>>>>
>>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>>>>>>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>>>>>>> _______________________________________________
>>>>>>> infinispan-dev mailing list
>>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>>
>>>>>> --
>>>>>> Manik Surtani
>>>>>> manik(a)jboss.org
>>>>>> Lead, Infinispan
>>>>>> Lead, JBoss Cache
>>>>>> http://www.infinispan.org
>>>>>> http://www.jbosscache.org
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> infinispan-dev mailing list
>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>>> --
>>>> Manik Surtani
>>>> manik(a)jboss.org
>>>> Lead, Infinispan
>>>> Lead, JBoss Cache
>>>> http://www.infinispan.org
>>>> http://www.jbosscache.org
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> --
>> Galder Zamarreño
>> Sr. Software Engineer
>> Infinispan, JBoss Cache
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Manik Surtani
> manik(a)jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
6:52 a.m.
On 09-10-13 5:54 AM, Mircea Markus wrote:
Thank you very much for sorting that out! I was really out of ideas
on
what to do next to make it work.
Kudos to you!
Hey, I just changed command line of Hudson to bind to a specific
interface and mcast address :)
Now we have to figure out why is jdbc cachestore almost always getting
stuck when we run udp - as it is now.
Cheers
2:52 p.m.
More surprises and good news. After I switched on bundling and
implemented https://jira.jboss.org/jira/browse/ISPN-192 test suite is
now running in 3 minutes and 36 seconds!!!
Have no idea why though, anyone?:)
On 09-10-13 7:52 AM, Vladimir Blagojevic wrote:
On 09-10-13 5:54 AM, Mircea Markus wrote:
> Thank you very much for sorting that out! I was really out of ideas on
> what to do next to make it work.
> Kudos to you!
>
>
Hey, I just changed command line of Hudson to bind to a specific
interface and mcast address :)
Now we have to figure out why is jdbc cachestore almost always getting
stuck when we run udp - as it is now.
Cheers
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
3:27 p.m.
On 09-10-14 3:52 PM, Vladimir Blagojevic wrote:
More surprises and good news. After I switched on bundling and
implemented https://jira.jboss.org/jira/browse/ISPN-192 test suite is
now running in 3 minutes and 36 seconds!!!
Have no idea why though, anyone?:)
I spoke too soon. We should allow bundling for async calls only - not
vice versa. It is hard to believe that bundling has such an effect on
running times....
5:19 a.m.
On 13 Oct 2009, at 04:12, Vladimir Blagojevic wrote:
UDP build is also up and running, no more problems for members to
find
each other. Hudson did not like binding to 127.0.0.1 for mcast
sockets....
Also does this mean ISPN-99 can be closed?
On 09-10-12 5:30 AM, Manik Surtani wrote:
> Great, thanks for looking into this!
>
> On 12 Oct 2009, at 06:20, Vladimir Blagojevic wrote:
>
>
>> On 09-10-11 5:53 AM, Manik Surtani wrote:
>>
>>> As long as it works and is non-intrusive, I'm happy. :-)
>>>
>> Done. There has been a regression in our test suite that I believe
>> is
>> unrelated to the change I made regarding these leaks. I did not pick
>> up
>> commits 936 and 937 and I do not see new failures [1]. So you guys
>> have
>> a look there. EhCache and jbc2InfinispanTransformerTest that Mircea
>> worked on pass if test suite is invoked from core directory but the
>> fail
>> if invoked from infinispan directory - as hudson does.
>>
>>
>> [1] http://hudson.jboss.org/hudson/job/Infinispan-trunk-JDK6-tcp/
>> 812/
>>
>>
>> http://fisheye.jboss.org/changelog/Infinispan/?cs=937
>> http://fisheye.jboss.org/changelog/Infinispan/?cs=936
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
> --
> Manik Surtani
> manik(a)jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org
5834
days inactive
5841
days old
infinispan-dev@lists.jboss.org
22 comments
6 participants
participants (6)
-
Bela Ban -
Galder Zamarreno -
Galder Zamarreno -
Manik Surtani -
Mircea Markus -
Vladimir Blagojevic