What is the problem with 5.0.1

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

What is the problem with 5.0.1

jszjsmith
Hello, I have been using 5.0.1 in our testing environment without issue and have recently installed 5.0.1 on production servers ready for deployment.

I have just seen today the message: "Because of a problem in the 5.0.1 release, we recommend you to use the 5.0.0 release instead!"

Before I roll back to 5.0.0 can somebody in the know explain the specifics of the problem so I can ascertain whether it is likely to affect us? I have tried to find the answer on JIRA but cannot pinpoint what the issue is.

Many thanks,
John.
Reply | Threaded
Open this post in threaded view
|

Re: What is the problem with 5.0.1

Gert Vanthienen
Administrator
L.S.,

While testing the release candidate for 5.1.0, we discovered a problem
when running even the basic quickstart examples.  While they run fine
at first glance, the context would remain "Creating" the second time
around.  The root cause appears to be in Karaf 2.3.5 - some changes
there are causing extra services to be registered in the OSGi Service
Registry while doing operations that interact with the JAAS
authentication service (e.g. doing authentication while opening a JMS
connection) - this causes those Blueprint extender threads to become
stuck, waiting for the global bundle lock to be released.  Because the
same version of Karaf is also being used in ServiceMix 5.0.1., that
has caused us to revert the main download page to point to 5.0.0.  We
are currently testing with a 5.1.0 build again which has rolled back
to Karaf 2.3.4 and from the tests I did this morning, things are
looking good there again.

You're definitely right that the banner needs a bit of clarification
and perhaps should sound a bit less "dramatic".  If you haven't bumped
into the problem while testing, I think you're probably fine for now.
With the 5.0.1 version of ServiceMix, the problem was pretty
consistent to reproduce.  If the tests with Karaf 2.3.4 are working
out fine, we will probably do a ServiceMix 5.1.0 release with that
version and then do a ServiceMix 5.1.1 release as soon as the problem
in Karaf is fixed in a new release.  Personally, I would be inclined
to to the same for 5.0.2, do a release there with the older version of
Karaf but the Camel/Activiti/... upgrades we had, and then follow up
with a 5.0.3 as soon as we have a Karaf fix.

About the banner on the 5.0.1 page, how about we rephrase that into
something a bit more descriptive like "Due to a problem in Apache
Karaf 2.3.5, you may experience problems using services that leverage
the JAAS authentication mechanism.  We therefore recommend users to
use Apache ServiceMix 5.0.0 instead."?

Regards,

Gert Vanthienen


On Thu, Jun 12, 2014 at 3:29 PM, jszjsmith <[hidden email]> wrote:

> Hello, I have been using 5.0.1 in our testing environment without issue and
> have recently installed 5.0.1 on production servers ready for deployment.
>
> I have just seen today the message: "Because of a problem in the 5.0.1
> release, we recommend you to use the 5.0.0 release instead!"
>
> Before I roll back to 5.0.0 can somebody in the know explain the specifics
> of the problem so I can ascertain whether it is likely to affect us? I have
> tried to find the answer on JIRA but cannot pinpoint what the issue is.
>
> Many thanks,
> John.
>
>
>
> --
> View this message in context: http://servicemix.396122.n5.nabble.com/What-is-the-problem-with-5-0-1-tp5720813.html
> Sent from the ServiceMix - Dev mailing list archive at Nabble.com.
Regards,

Gert Vanthienen
------------------------
Open Source SOA: http://fusesource.com
Blog: http://gertvanthienen.blogspot.com/
Reply | Threaded
Open this post in threaded view
|

Re: What is the problem with 5.0.1

Jean-Baptiste Onofré
Hi Gert,

I'm agree to update the message: it's more factual and less dramatic.

On my side, I'm moving forward to release Karaf 2.3.6 fixing the issue
introduced in 2.3.5. As I said on the Karaf mailing list, 2.3.6 should
be in vote next week.

Regards
JB

On 06/13/2014 01:03 PM, Gert Vanthienen wrote:

> L.S.,
>
> While testing the release candidate for 5.1.0, we discovered a problem
> when running even the basic quickstart examples.  While they run fine
> at first glance, the context would remain "Creating" the second time
> around.  The root cause appears to be in Karaf 2.3.5 - some changes
> there are causing extra services to be registered in the OSGi Service
> Registry while doing operations that interact with the JAAS
> authentication service (e.g. doing authentication while opening a JMS
> connection) - this causes those Blueprint extender threads to become
> stuck, waiting for the global bundle lock to be released.  Because the
> same version of Karaf is also being used in ServiceMix 5.0.1., that
> has caused us to revert the main download page to point to 5.0.0.  We
> are currently testing with a 5.1.0 build again which has rolled back
> to Karaf 2.3.4 and from the tests I did this morning, things are
> looking good there again.
>
> You're definitely right that the banner needs a bit of clarification
> and perhaps should sound a bit less "dramatic".  If you haven't bumped
> into the problem while testing, I think you're probably fine for now.
> With the 5.0.1 version of ServiceMix, the problem was pretty
> consistent to reproduce.  If the tests with Karaf 2.3.4 are working
> out fine, we will probably do a ServiceMix 5.1.0 release with that
> version and then do a ServiceMix 5.1.1 release as soon as the problem
> in Karaf is fixed in a new release.  Personally, I would be inclined
> to to the same for 5.0.2, do a release there with the older version of
> Karaf but the Camel/Activiti/... upgrades we had, and then follow up
> with a 5.0.3 as soon as we have a Karaf fix.
>
> About the banner on the 5.0.1 page, how about we rephrase that into
> something a bit more descriptive like "Due to a problem in Apache
> Karaf 2.3.5, you may experience problems using services that leverage
> the JAAS authentication mechanism.  We therefore recommend users to
> use Apache ServiceMix 5.0.0 instead."?
>
> Regards,
>
> Gert Vanthienen
>
>
> On Thu, Jun 12, 2014 at 3:29 PM, jszjsmith <[hidden email]> wrote:
>> Hello, I have been using 5.0.1 in our testing environment without issue and
>> have recently installed 5.0.1 on production servers ready for deployment.
>>
>> I have just seen today the message: "Because of a problem in the 5.0.1
>> release, we recommend you to use the 5.0.0 release instead!"
>>
>> Before I roll back to 5.0.0 can somebody in the know explain the specifics
>> of the problem so I can ascertain whether it is likely to affect us? I have
>> tried to find the answer on JIRA but cannot pinpoint what the issue is.
>>
>> Many thanks,
>> John.
>>
>>
>>
>> --
>> View this message in context: http://servicemix.396122.n5.nabble.com/What-is-the-problem-with-5-0-1-tp5720813.html
>> Sent from the ServiceMix - Dev mailing list archive at Nabble.com.

--
Jean-Baptiste Onofré
[hidden email]
http://blog.nanthrax.net
Talend - http://www.talend.com
Reply | Threaded
Open this post in threaded view
|

Re: What is the problem with 5.0.1

Raúl Kripalani
In reply to this post by Gert Vanthienen
Many thanks for the explanation, Gert!

*Raúl Kripalani*
Apache Camel PMC Member & Committer | Enterprise Architect, Open Source
Integration specialist
http://about.me/raulkripalani | http://www.linkedin.com/in/raulkripalani
http://blog.raulkr.net | twitter: @raulvk

On Fri, Jun 13, 2014 at 12:03 PM, Gert Vanthienen <[hidden email]
> wrote:

> L.S.,
>
> While testing the release candidate for 5.1.0, we discovered a problem
> when running even the basic quickstart examples.  While they run fine
> at first glance, the context would remain "Creating" the second time
> around.  The root cause appears to be in Karaf 2.3.5 - some changes
> there are causing extra services to be registered in the OSGi Service
> Registry while doing operations that interact with the JAAS
> authentication service (e.g. doing authentication while opening a JMS
> connection) - this causes those Blueprint extender threads to become
> stuck, waiting for the global bundle lock to be released.  Because the
> same version of Karaf is also being used in ServiceMix 5.0.1., that
> has caused us to revert the main download page to point to 5.0.0.  We
> are currently testing with a 5.1.0 build again which has rolled back
> to Karaf 2.3.4 and from the tests I did this morning, things are
> looking good there again.
>
> You're definitely right that the banner needs a bit of clarification
> and perhaps should sound a bit less "dramatic".  If you haven't bumped
> into the problem while testing, I think you're probably fine for now.
> With the 5.0.1 version of ServiceMix, the problem was pretty
> consistent to reproduce.  If the tests with Karaf 2.3.4 are working
> out fine, we will probably do a ServiceMix 5.1.0 release with that
> version and then do a ServiceMix 5.1.1 release as soon as the problem
> in Karaf is fixed in a new release.  Personally, I would be inclined
> to to the same for 5.0.2, do a release there with the older version of
> Karaf but the Camel/Activiti/... upgrades we had, and then follow up
> with a 5.0.3 as soon as we have a Karaf fix.
>
> About the banner on the 5.0.1 page, how about we rephrase that into
> something a bit more descriptive like "Due to a problem in Apache
> Karaf 2.3.5, you may experience problems using services that leverage
> the JAAS authentication mechanism.  We therefore recommend users to
> use Apache ServiceMix 5.0.0 instead."?
>
> Regards,
>
> Gert Vanthienen
>
>
> On Thu, Jun 12, 2014 at 3:29 PM, jszjsmith <[hidden email]>
> wrote:
> > Hello, I have been using 5.0.1 in our testing environment without issue
> and
> > have recently installed 5.0.1 on production servers ready for deployment.
> >
> > I have just seen today the message: "Because of a problem in the 5.0.1
> > release, we recommend you to use the 5.0.0 release instead!"
> >
> > Before I roll back to 5.0.0 can somebody in the know explain the
> specifics
> > of the problem so I can ascertain whether it is likely to affect us? I
> have
> > tried to find the answer on JIRA but cannot pinpoint what the issue is.
> >
> > Many thanks,
> > John.
> >
> >
> >
> > --
> > View this message in context:
> http://servicemix.396122.n5.nabble.com/What-is-the-problem-with-5-0-1-tp5720813.html
> > Sent from the ServiceMix - Dev mailing list archive at Nabble.com.
>