New Branch SUREFIRE-1342

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

New Branch SUREFIRE-1342

Tibor Digana
Hi All,

The new branch SUREFIRE-1342 solves an issue when entire testset completed
however the surefire's forked JVM  finished printing serious issue. We know
that the JVM did not crash however it looks so:

The forked VM terminated without saying properly goodbye. VM crash or
System.exit called ?


The problem is solved in the branch and ready to be tested.

The entire problem was concurrency issue where the forked JVM sent "bye"
event to the Maven process via stdout but Maven process has not drained
shared memory yet and Maven process was therefore slower to receive the
event than the forked process which exited. Due to the "bye" event was not
determined by Maven process in particular time, this error came up.
We implemented ACK command which confirms such event has been received by
Maven process. The shared memory is drained directly by Maven process.

Cheers
Tibor
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

gboue
Hi,

I finished setting up a FreeBSD VM with a couple of Maven and JDK
versions, so I'll be testing this extensively (on Ubuntu as well).
Thanks for the work!

Guillaume


Le 13/03/2017 à 10:46, Tibor Digana a écrit :

> Hi All,
>
> The new branch SUREFIRE-1342 solves an issue when entire testset completed
> however the surefire's forked JVM  finished printing serious issue. We know
> that the JVM did not crash however it looks so:
>
> The forked VM terminated without saying properly goodbye. VM crash or
> System.exit called ?
>
>
> The problem is solved in the branch and ready to be tested.
>
> The entire problem was concurrency issue where the forked JVM sent "bye"
> event to the Maven process via stdout but Maven process has not drained
> shared memory yet and Maven process was therefore slower to receive the
> event than the forked process which exited. Due to the "bye" event was not
> determined by Maven process in particular time, this error came up.
> We implemented ACK command which confirms such event has been received by
> Maven process. The shared memory is drained directly by Maven process.
>
> Cheers
> Tibor
>


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

Tibor Digana-2
We have performance tests. This means your machine should run only this
build. The GC may affect the success therefore my fine tuning is
MAVEN_OPTC=-server -Xmx768m -XX:MaxMetaspaceSize=864m.
Of course MaxMetaspaceSize is available in JDK8. Use MaxPemSize in JDK7 or
JDK6.
The ideal would be to have 4 Core CPU x64, SSD where the build takes only
45 minutes.
$ mvn -P run-its install

On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]> wrote:

> Hi,
>
> I finished setting up a FreeBSD VM with a couple of Maven and JDK
> versions, so I'll be testing this extensively (on Ubuntu as well). Thanks
> for the work!
>
> Guillaume
>
>
>
> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>
>> Hi All,
>>
>> The new branch SUREFIRE-1342 solves an issue when entire testset completed
>> however the surefire's forked JVM  finished printing serious issue. We
>> know
>> that the JVM did not crash however it looks so:
>>
>> The forked VM terminated without saying properly goodbye. VM crash or
>> System.exit called ?
>>
>>
>> The problem is solved in the branch and ready to be tested.
>>
>> The entire problem was concurrency issue where the forked JVM sent "bye"
>> event to the Maven process via stdout but Maven process has not drained
>> shared memory yet and Maven process was therefore slower to receive the
>> event than the forked process which exited. Due to the "bye" event was not
>> determined by Maven process in particular time, this error came up.
>> We implemented ACK command which confirms such event has been received by
>> Maven process. The shared memory is drained directly by Maven process.
>>
>> Cheers
>> Tibor
>>
>>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>


--
Cheers
Tibor
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

Tibor Digana
In reply to this post by gboue
Hi Guillaume Boué,

Have you found a spare time to test the branch?

Cheers
Tibor

On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]> wrote:

> Hi,
>
> I finished setting up a FreeBSD VM with a couple of Maven and JDK
> versions, so I'll be testing this extensively (on Ubuntu as well). Thanks
> for the work!
>
> Guillaume
>
>
>
> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>
>> Hi All,
>>
>> The new branch SUREFIRE-1342 solves an issue when entire testset completed
>> however the surefire's forked JVM  finished printing serious issue. We
>> know
>> that the JVM did not crash however it looks so:
>>
>> The forked VM terminated without saying properly goodbye. VM crash or
>> System.exit called ?
>>
>>
>> The problem is solved in the branch and ready to be tested.
>>
>> The entire problem was concurrency issue where the forked JVM sent "bye"
>> event to the Maven process via stdout but Maven process has not drained
>> shared memory yet and Maven process was therefore slower to receive the
>> event than the forked process which exited. Due to the "bye" event was not
>> determined by Maven process in particular time, this error came up.
>> We implemented ACK command which confirms such event has been received by
>> Maven process. The shared memory is drained directly by Maven process.
>>
>> Cheers
>> Tibor
>>
>>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

gboue
Yes, I finished running the tests multiple times on FreeBSD with Maven
3.5.0-alpha-1, and they are all passing! The tests were however
painfully slow (more than 4 hours), it might be an issue with my VM (64
bits, only has 1 Go of RAM and 1 vCPU), I'll see how much time it takes
with those same characteristics, but on Ubuntu.

I will be running them again on Windows and Ubuntu with JDK 7 and 8 over
the next few days.

Guillaume


Le 15/03/2017 à 21:40, Tibor Digana a écrit :

> Hi Guillaume Boué,
>
> Have you found a spare time to test the branch?
>
> Cheers
> Tibor
>
> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]> wrote:
>
>> Hi,
>>
>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>> versions, so I'll be testing this extensively (on Ubuntu as well). Thanks
>> for the work!
>>
>> Guillaume
>>
>>
>>
>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>
>>> Hi All,
>>>
>>> The new branch SUREFIRE-1342 solves an issue when entire testset completed
>>> however the surefire's forked JVM  finished printing serious issue. We
>>> know
>>> that the JVM did not crash however it looks so:
>>>
>>> The forked VM terminated without saying properly goodbye. VM crash or
>>> System.exit called ?
>>>
>>>
>>> The problem is solved in the branch and ready to be tested.
>>>
>>> The entire problem was concurrency issue where the forked JVM sent "bye"
>>> event to the Maven process via stdout but Maven process has not drained
>>> shared memory yet and Maven process was therefore slower to receive the
>>> event than the forked process which exited. Due to the "bye" event was not
>>> determined by Maven process in particular time, this error came up.
>>> We implemented ACK command which confirms such event has been received by
>>> Maven process. The shared memory is drained directly by Maven process.
>>>
>>> Cheers
>>> Tibor
>>>
>>>
>> ---
>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>> logiciel antivirus Avast.
>> https://www.avast.com/antivirus
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

Tibor Digana-2
Yes, there are a lot of tests 1200 altogether. For instance I have 4 Cores
of CPU and the build takes 45 minutes. In your case 1 Core of Virtual CPU
which makes the difference. Tuning of JVM is also important. Not too much
and not too less of Xmx cca 700 MB and the same or more for permanent part
of memory. I think JVM does not run in server mode if RAM < 2GiB even if
you have x64.

On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]> wrote:

> Yes, I finished running the tests multiple times on FreeBSD with Maven
> 3.5.0-alpha-1, and they are all passing! The tests were however painfully
> slow (more than 4 hours), it might be an issue with my VM (64 bits, only
> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
> same characteristics, but on Ubuntu.
>
> I will be running them again on Windows and Ubuntu with JDK 7 and 8 over
> the next few days.
>
> Guillaume
>
>
>
> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>
>> Hi Guillaume Boué,
>>
>> Have you found a spare time to test the branch?
>>
>> Cheers
>> Tibor
>>
>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>> wrote:
>>
>> Hi,
>>>
>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>> versions, so I'll be testing this extensively (on Ubuntu as well). Thanks
>>> for the work!
>>>
>>> Guillaume
>>>
>>>
>>>
>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>
>>> Hi All,
>>>>
>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>> completed
>>>> however the surefire's forked JVM  finished printing serious issue. We
>>>> know
>>>> that the JVM did not crash however it looks so:
>>>>
>>>> The forked VM terminated without saying properly goodbye. VM crash or
>>>> System.exit called ?
>>>>
>>>>
>>>> The problem is solved in the branch and ready to be tested.
>>>>
>>>> The entire problem was concurrency issue where the forked JVM sent "bye"
>>>> event to the Maven process via stdout but Maven process has not drained
>>>> shared memory yet and Maven process was therefore slower to receive the
>>>> event than the forked process which exited. Due to the "bye" event was
>>>> not
>>>> determined by Maven process in particular time, this error came up.
>>>> We implemented ACK command which confirms such event has been received
>>>> by
>>>> Maven process. The shared memory is drained directly by Maven process.
>>>>
>>>> Cheers
>>>> Tibor
>>>>
>>>>
>>>> ---
>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>> logiciel antivirus Avast.
>>> https://www.avast.com/antivirus
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>


--
Cheers
Tibor
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

gboue
So with JDK 8 (64 bit where possible), the tests are all passing on the
systems I tested: FreeBSD 11.0 x64, Ubuntu 16.04 (x32 and x64), and
Windows 10 x64.

I have two tests consistently in error with OpenJDK 7 (and Maven
3.5.0-alpha-1), at least on Ubuntu:

Failed tests:
invokeRuntimeException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT):
expecting non-empty, but it was empty
invokeReporterException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT):
expecting non-empty, but it was empty

They are also failing when ran individually if passing
-Dit.test=Surefire141PluggableProvidersIT to the Maven command.

I will test with a JDK 6 right now.


Le 17/03/2017 à 06:12, Tibor Digana a écrit :

> Yes, there are a lot of tests 1200 altogether. For instance I have 4 Cores
> of CPU and the build takes 45 minutes. In your case 1 Core of Virtual CPU
> which makes the difference. Tuning of JVM is also important. Not too much
> and not too less of Xmx cca 700 MB and the same or more for permanent part
> of memory. I think JVM does not run in server mode if RAM < 2GiB even if
> you have x64.
>
> On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]> wrote:
>
>> Yes, I finished running the tests multiple times on FreeBSD with Maven
>> 3.5.0-alpha-1, and they are all passing! The tests were however painfully
>> slow (more than 4 hours), it might be an issue with my VM (64 bits, only
>> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
>> same characteristics, but on Ubuntu.
>>
>> I will be running them again on Windows and Ubuntu with JDK 7 and 8 over
>> the next few days.
>>
>> Guillaume
>>
>>
>>
>> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>>
>>> Hi Guillaume Boué,
>>>
>>> Have you found a spare time to test the branch?
>>>
>>> Cheers
>>> Tibor
>>>
>>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>>> wrote:
>>>
>>> Hi,
>>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>>> versions, so I'll be testing this extensively (on Ubuntu as well). Thanks
>>>> for the work!
>>>>
>>>> Guillaume
>>>>
>>>>
>>>>
>>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>>
>>>> Hi All,
>>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>>> completed
>>>>> however the surefire's forked JVM  finished printing serious issue. We
>>>>> know
>>>>> that the JVM did not crash however it looks so:
>>>>>
>>>>> The forked VM terminated without saying properly goodbye. VM crash or
>>>>> System.exit called ?
>>>>>
>>>>>
>>>>> The problem is solved in the branch and ready to be tested.
>>>>>
>>>>> The entire problem was concurrency issue where the forked JVM sent "bye"
>>>>> event to the Maven process via stdout but Maven process has not drained
>>>>> shared memory yet and Maven process was therefore slower to receive the
>>>>> event than the forked process which exited. Due to the "bye" event was
>>>>> not
>>>>> determined by Maven process in particular time, this error came up.
>>>>> We implemented ACK command which confirms such event has been received
>>>>> by
>>>>> Maven process. The shared memory is drained directly by Maven process.
>>>>>
>>>>> Cheers
>>>>> Tibor
>>>>>
>>>>>
>>>>> ---
>>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>>> logiciel antivirus Avast.
>>>> https://www.avast.com/antivirus
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>> ---
>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>> logiciel antivirus Avast.
>> https://www.avast.com/antivirus
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

gboue


Le 18/03/2017 à 20:49, Guillaume Boué a écrit :

> So with JDK 8 (64 bit where possible), the tests are all passing on
> the systems I tested: FreeBSD 11.0 x64, Ubuntu 16.04 (x32 and x64),
> and Windows 10 x64.
>
> I have two tests consistently in error with OpenJDK 7 (and Maven
> 3.5.0-alpha-1), at least on Ubuntu:
>
> Failed tests:
> invokeRuntimeException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT):
> expecting non-empty, but it was empty
> invokeReporterException(org.apache.maven.surefire.its.jiras.Surefire141PluggableProvidersIT):
> expecting non-empty, but it was empty
>
> They are also failing when ran individually if passing
> -Dit.test=Surefire141PluggableProvidersIT to the Maven command.
>
> I will test with a JDK 6 right now.

Same errors with Oracle JDK 6, and Maven 3.2.5.

>
>
> Le 17/03/2017 à 06:12, Tibor Digana a écrit :
>> Yes, there are a lot of tests 1200 altogether. For instance I have 4
>> Cores
>> of CPU and the build takes 45 minutes. In your case 1 Core of Virtual
>> CPU
>> which makes the difference. Tuning of JVM is also important. Not too
>> much
>> and not too less of Xmx cca 700 MB and the same or more for permanent
>> part
>> of memory. I think JVM does not run in server mode if RAM < 2GiB even if
>> you have x64.
>>
>> On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]>
>> wrote:
>>
>>> Yes, I finished running the tests multiple times on FreeBSD with Maven
>>> 3.5.0-alpha-1, and they are all passing! The tests were however
>>> painfully
>>> slow (more than 4 hours), it might be an issue with my VM (64 bits,
>>> only
>>> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
>>> same characteristics, but on Ubuntu.
>>>
>>> I will be running them again on Windows and Ubuntu with JDK 7 and 8
>>> over
>>> the next few days.
>>>
>>> Guillaume
>>>
>>>
>>>
>>> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>>>
>>>> Hi Guillaume Boué,
>>>>
>>>> Have you found a spare time to test the branch?
>>>>
>>>> Cheers
>>>> Tibor
>>>>
>>>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>>>> wrote:
>>>>
>>>> Hi,
>>>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>>>> versions, so I'll be testing this extensively (on Ubuntu as well).
>>>>> Thanks
>>>>> for the work!
>>>>>
>>>>> Guillaume
>>>>>
>>>>>
>>>>>
>>>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>>>
>>>>> Hi All,
>>>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>>>> completed
>>>>>> however the surefire's forked JVM  finished printing serious
>>>>>> issue. We
>>>>>> know
>>>>>> that the JVM did not crash however it looks so:
>>>>>>
>>>>>> The forked VM terminated without saying properly goodbye. VM
>>>>>> crash or
>>>>>> System.exit called ?
>>>>>>
>>>>>>
>>>>>> The problem is solved in the branch and ready to be tested.
>>>>>>
>>>>>> The entire problem was concurrency issue where the forked JVM
>>>>>> sent "bye"
>>>>>> event to the Maven process via stdout but Maven process has not
>>>>>> drained
>>>>>> shared memory yet and Maven process was therefore slower to
>>>>>> receive the
>>>>>> event than the forked process which exited. Due to the "bye"
>>>>>> event was
>>>>>> not
>>>>>> determined by Maven process in particular time, this error came up.
>>>>>> We implemented ACK command which confirms such event has been
>>>>>> received
>>>>>> by
>>>>>> Maven process. The shared memory is drained directly by Maven
>>>>>> process.
>>>>>>
>>>>>> Cheers
>>>>>> Tibor
>>>>>>
>>>>>>
>>>>>> ---
>>>>> L'absence de virus dans ce courrier électronique a été vérifiée
>>>>> par le
>>>>> logiciel antivirus Avast.
>>>>> https://www.avast.com/antivirus
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>>
>>> ---
>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>> logiciel antivirus Avast.
>>> https://www.avast.com/antivirus
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>
>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

Tibor Digana-2
In reply to this post by gboue
Please send the dump files from
surefire-integration-tests/target/Surefire141PluggableProvidersIT_invokeRuntimeException
surefire-integration-tests/target/Surefire141PluggableProvidersIT_
invokeReporterException

They must exist in both.

This issue is also concurrency issue but since JVM exit code  is 1, the ACK
is not applicable.

I am thinking about two assertions statements instead of current one.

On Sat, Mar 18, 2017 at 8:49 PM, Guillaume Boué <[hidden email]> wrote:

> So with JDK 8 (64 bit where possible), the tests are all passing on the
> systems I tested: FreeBSD 11.0 x64, Ubuntu 16.04 (x32 and x64), and Windows
> 10 x64.
>
> I have two tests consistently in error with OpenJDK 7 (and Maven
> 3.5.0-alpha-1), at least on Ubuntu:
>
> Failed tests:
> invokeRuntimeException(org.apache.maven.surefire.its.jiras.S
> urefire141PluggableProvidersIT): expecting non-empty, but it was empty
> invokeReporterException(org.apache.maven.surefire.its.jiras.
> Surefire141PluggableProvidersIT): expecting non-empty, but it was empty
>
> They are also failing when ran individually if passing
> -Dit.test=Surefire141PluggableProvidersIT to the Maven command.
>
> I will test with a JDK 6 right now.
>
>
> Le 17/03/2017 à 06:12, Tibor Digana a écrit :
>
>> Yes, there are a lot of tests 1200 altogether. For instance I have 4 Cores
>> of CPU and the build takes 45 minutes. In your case 1 Core of Virtual CPU
>> which makes the difference. Tuning of JVM is also important. Not too much
>> and not too less of Xmx cca 700 MB and the same or more for permanent part
>> of memory. I think JVM does not run in server mode if RAM < 2GiB even if
>> you have x64.
>>
>> On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]>
>> wrote:
>>
>> Yes, I finished running the tests multiple times on FreeBSD with Maven
>>> 3.5.0-alpha-1, and they are all passing! The tests were however painfully
>>> slow (more than 4 hours), it might be an issue with my VM (64 bits, only
>>> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
>>> same characteristics, but on Ubuntu.
>>>
>>> I will be running them again on Windows and Ubuntu with JDK 7 and 8 over
>>> the next few days.
>>>
>>> Guillaume
>>>
>>>
>>>
>>> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>>>
>>> Hi Guillaume Boué,
>>>>
>>>> Have you found a spare time to test the branch?
>>>>
>>>> Cheers
>>>> Tibor
>>>>
>>>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>>>> wrote:
>>>>
>>>> Hi,
>>>>
>>>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>>>> versions, so I'll be testing this extensively (on Ubuntu as well).
>>>>> Thanks
>>>>> for the work!
>>>>>
>>>>> Guillaume
>>>>>
>>>>>
>>>>>
>>>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>>>
>>>>> Hi All,
>>>>>
>>>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>>>> completed
>>>>>> however the surefire's forked JVM  finished printing serious issue. We
>>>>>> know
>>>>>> that the JVM did not crash however it looks so:
>>>>>>
>>>>>> The forked VM terminated without saying properly goodbye. VM crash or
>>>>>> System.exit called ?
>>>>>>
>>>>>>
>>>>>> The problem is solved in the branch and ready to be tested.
>>>>>>
>>>>>> The entire problem was concurrency issue where the forked JVM sent
>>>>>> "bye"
>>>>>> event to the Maven process via stdout but Maven process has not
>>>>>> drained
>>>>>> shared memory yet and Maven process was therefore slower to receive
>>>>>> the
>>>>>> event than the forked process which exited. Due to the "bye" event was
>>>>>> not
>>>>>> determined by Maven process in particular time, this error came up.
>>>>>> We implemented ACK command which confirms such event has been received
>>>>>> by
>>>>>> Maven process. The shared memory is drained directly by Maven process.
>>>>>>
>>>>>> Cheers
>>>>>> Tibor
>>>>>>
>>>>>>
>>>>>> ---
>>>>>>
>>>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>>>> logiciel antivirus Avast.
>>>>> https://www.avast.com/antivirus
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>>
>>>>> ---
>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>> logiciel antivirus Avast.
>>> https://www.avast.com/antivirus
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>


--
Cheers
Tibor
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

gboue
This is probably the problem, I don't have any dump files. I've attached
the log.txt files for both tests.

It doesn't seem to be related to the branch though. They work fine in
2.19.1 and also fail with current master.


Le 18/03/2017 à 21:57, Tibor Digana a écrit :

> Please send the dump files from
> surefire-integration-tests/target/Surefire141PluggableProvidersIT_invokeRuntimeException
> surefire-integration-tests/target/Surefire141PluggableProvidersIT_
> invokeReporterException
>
> They must exist in both.
>
> This issue is also concurrency issue but since JVM exit code  is 1, the ACK
> is not applicable.
>
> I am thinking about two assertions statements instead of current one.
>
> On Sat, Mar 18, 2017 at 8:49 PM, Guillaume Boué <[hidden email]> wrote:
>
>> So with JDK 8 (64 bit where possible), the tests are all passing on the
>> systems I tested: FreeBSD 11.0 x64, Ubuntu 16.04 (x32 and x64), and Windows
>> 10 x64.
>>
>> I have two tests consistently in error with OpenJDK 7 (and Maven
>> 3.5.0-alpha-1), at least on Ubuntu:
>>
>> Failed tests:
>> invokeRuntimeException(org.apache.maven.surefire.its.jiras.S
>> urefire141PluggableProvidersIT): expecting non-empty, but it was empty
>> invokeReporterException(org.apache.maven.surefire.its.jiras.
>> Surefire141PluggableProvidersIT): expecting non-empty, but it was empty
>>
>> They are also failing when ran individually if passing
>> -Dit.test=Surefire141PluggableProvidersIT to the Maven command.
>>
>> I will test with a JDK 6 right now.
>>
>>
>> Le 17/03/2017 à 06:12, Tibor Digana a écrit :
>>
>>> Yes, there are a lot of tests 1200 altogether. For instance I have 4 Cores
>>> of CPU and the build takes 45 minutes. In your case 1 Core of Virtual CPU
>>> which makes the difference. Tuning of JVM is also important. Not too much
>>> and not too less of Xmx cca 700 MB and the same or more for permanent part
>>> of memory. I think JVM does not run in server mode if RAM < 2GiB even if
>>> you have x64.
>>>
>>> On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]>
>>> wrote:
>>>
>>> Yes, I finished running the tests multiple times on FreeBSD with Maven
>>>> 3.5.0-alpha-1, and they are all passing! The tests were however painfully
>>>> slow (more than 4 hours), it might be an issue with my VM (64 bits, only
>>>> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
>>>> same characteristics, but on Ubuntu.
>>>>
>>>> I will be running them again on Windows and Ubuntu with JDK 7 and 8 over
>>>> the next few days.
>>>>
>>>> Guillaume
>>>>
>>>>
>>>>
>>>> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>>>>
>>>> Hi Guillaume Boué,
>>>>> Have you found a spare time to test the branch?
>>>>>
>>>>> Cheers
>>>>> Tibor
>>>>>
>>>>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>>>>> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>>>>> versions, so I'll be testing this extensively (on Ubuntu as well).
>>>>>> Thanks
>>>>>> for the work!
>>>>>>
>>>>>> Guillaume
>>>>>>
>>>>>>
>>>>>>
>>>>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>>>>
>>>>>> Hi All,
>>>>>>
>>>>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>>>>> completed
>>>>>>> however the surefire's forked JVM  finished printing serious issue. We
>>>>>>> know
>>>>>>> that the JVM did not crash however it looks so:
>>>>>>>
>>>>>>> The forked VM terminated without saying properly goodbye. VM crash or
>>>>>>> System.exit called ?
>>>>>>>
>>>>>>>
>>>>>>> The problem is solved in the branch and ready to be tested.
>>>>>>>
>>>>>>> The entire problem was concurrency issue where the forked JVM sent
>>>>>>> "bye"
>>>>>>> event to the Maven process via stdout but Maven process has not
>>>>>>> drained
>>>>>>> shared memory yet and Maven process was therefore slower to receive
>>>>>>> the
>>>>>>> event than the forked process which exited. Due to the "bye" event was
>>>>>>> not
>>>>>>> determined by Maven process in particular time, this error came up.
>>>>>>> We implemented ACK command which confirms such event has been received
>>>>>>> by
>>>>>>> Maven process. The shared memory is drained directly by Maven process.
>>>>>>>
>>>>>>> Cheers
>>>>>>> Tibor
>>>>>>>
>>>>>>>
>>>>>>> ---
>>>>>>>
>>>>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>>>>> logiciel antivirus Avast.
>>>>>> https://www.avast.com/antivirus
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: [hidden email]
>>>>>> For additional commands, e-mail: [hidden email]
>>>>>>
>>>>>>
>>>>>>
>>>>>> ---
>>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>>> logiciel antivirus Avast.
>>>> https://www.avast.com/antivirus
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>> ---
>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>> logiciel antivirus Avast.
>> https://www.avast.com/antivirus
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

invokeRuntimeException_log.txt (22K) Download Attachment
invokeReporterException_log.txt (22K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: New Branch SUREFIRE-1342

Tibor Digana-2
I thought you was testing the branch SUREFIRE-1342.
git clone -b SUREFIRE-1342

On Sat, Mar 18, 2017 at 10:48 PM, Guillaume Boué <[hidden email]> wrote:

> This is probably the problem, I don't have any dump files. I've attached
> the log.txt files for both tests.
>
> It doesn't seem to be related to the branch though. They work fine in
> 2.19.1 and also fail with current master.
>
>
>
> Le 18/03/2017 à 21:57, Tibor Digana a écrit :
>
>> Please send the dump files from
>> surefire-integration-tests/target/Surefire141PluggableProvid
>> ersIT_invokeRuntimeException
>> surefire-integration-tests/target/Surefire141PluggableProvidersIT_
>> invokeReporterException
>>
>> They must exist in both.
>>
>> This issue is also concurrency issue but since JVM exit code  is 1, the
>> ACK
>> is not applicable.
>>
>> I am thinking about two assertions statements instead of current one.
>>
>> On Sat, Mar 18, 2017 at 8:49 PM, Guillaume Boué <[hidden email]> wrote:
>>
>> So with JDK 8 (64 bit where possible), the tests are all passing on the
>>> systems I tested: FreeBSD 11.0 x64, Ubuntu 16.04 (x32 and x64), and
>>> Windows
>>> 10 x64.
>>>
>>> I have two tests consistently in error with OpenJDK 7 (and Maven
>>> 3.5.0-alpha-1), at least on Ubuntu:
>>>
>>> Failed tests:
>>> invokeRuntimeException(org.apache.maven.surefire.its.jiras.S
>>> urefire141PluggableProvidersIT): expecting non-empty, but it was empty
>>> invokeReporterException(org.apache.maven.surefire.its.jiras.
>>> Surefire141PluggableProvidersIT): expecting non-empty, but it was empty
>>>
>>> They are also failing when ran individually if passing
>>> -Dit.test=Surefire141PluggableProvidersIT to the Maven command.
>>>
>>> I will test with a JDK 6 right now.
>>>
>>>
>>> Le 17/03/2017 à 06:12, Tibor Digana a écrit :
>>>
>>> Yes, there are a lot of tests 1200 altogether. For instance I have 4
>>>> Cores
>>>> of CPU and the build takes 45 minutes. In your case 1 Core of Virtual
>>>> CPU
>>>> which makes the difference. Tuning of JVM is also important. Not too
>>>> much
>>>> and not too less of Xmx cca 700 MB and the same or more for permanent
>>>> part
>>>> of memory. I think JVM does not run in server mode if RAM < 2GiB even if
>>>> you have x64.
>>>>
>>>> On Fri, Mar 17, 2017 at 12:53 AM, Guillaume Boué <[hidden email]>
>>>> wrote:
>>>>
>>>> Yes, I finished running the tests multiple times on FreeBSD with Maven
>>>>
>>>>> 3.5.0-alpha-1, and they are all passing! The tests were however
>>>>> painfully
>>>>> slow (more than 4 hours), it might be an issue with my VM (64 bits,
>>>>> only
>>>>> has 1 Go of RAM and 1 vCPU), I'll see how much time it takes with those
>>>>> same characteristics, but on Ubuntu.
>>>>>
>>>>> I will be running them again on Windows and Ubuntu with JDK 7 and 8
>>>>> over
>>>>> the next few days.
>>>>>
>>>>> Guillaume
>>>>>
>>>>>
>>>>>
>>>>> Le 15/03/2017 à 21:40, Tibor Digana a écrit :
>>>>>
>>>>> Hi Guillaume Boué,
>>>>>
>>>>>> Have you found a spare time to test the branch?
>>>>>>
>>>>>> Cheers
>>>>>> Tibor
>>>>>>
>>>>>> On Mon, Mar 13, 2017 at 11:04 PM, Guillaume Boué <[hidden email]>
>>>>>> wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I finished setting up a FreeBSD VM with a couple of Maven and JDK
>>>>>>> versions, so I'll be testing this extensively (on Ubuntu as well).
>>>>>>> Thanks
>>>>>>> for the work!
>>>>>>>
>>>>>>> Guillaume
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Le 13/03/2017 à 10:46, Tibor Digana a écrit :
>>>>>>>
>>>>>>> Hi All,
>>>>>>>
>>>>>>> The new branch SUREFIRE-1342 solves an issue when entire testset
>>>>>>>> completed
>>>>>>>> however the surefire's forked JVM  finished printing serious issue.
>>>>>>>> We
>>>>>>>> know
>>>>>>>> that the JVM did not crash however it looks so:
>>>>>>>>
>>>>>>>> The forked VM terminated without saying properly goodbye. VM crash
>>>>>>>> or
>>>>>>>> System.exit called ?
>>>>>>>>
>>>>>>>>
>>>>>>>> The problem is solved in the branch and ready to be tested.
>>>>>>>>
>>>>>>>> The entire problem was concurrency issue where the forked JVM sent
>>>>>>>> "bye"
>>>>>>>> event to the Maven process via stdout but Maven process has not
>>>>>>>> drained
>>>>>>>> shared memory yet and Maven process was therefore slower to receive
>>>>>>>> the
>>>>>>>> event than the forked process which exited. Due to the "bye" event
>>>>>>>> was
>>>>>>>> not
>>>>>>>> determined by Maven process in particular time, this error came up.
>>>>>>>> We implemented ACK command which confirms such event has been
>>>>>>>> received
>>>>>>>> by
>>>>>>>> Maven process. The shared memory is drained directly by Maven
>>>>>>>> process.
>>>>>>>>
>>>>>>>> Cheers
>>>>>>>> Tibor
>>>>>>>>
>>>>>>>>
>>>>>>>> ---
>>>>>>>>
>>>>>>>> L'absence de virus dans ce courrier électronique a été vérifiée par
>>>>>>> le
>>>>>>> logiciel antivirus Avast.
>>>>>>> https://www.avast.com/antivirus
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------
>>>>>>> ---------
>>>>>>> To unsubscribe, e-mail: [hidden email]
>>>>>>> For additional commands, e-mail: [hidden email]
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ---
>>>>>>>
>>>>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>>>> logiciel antivirus Avast.
>>>>> https://www.avast.com/antivirus
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>>
>>>>> ---
>>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>>> logiciel antivirus Avast.
>>> https://www.avast.com/antivirus
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>>
>
>
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>



--
Cheers
Tibor
Loading...