11:47 a.m.
Hi.
I'm going to remove the BlueOcean plugin from Ci.
Currently, when we receive an email because a build is failed, the
link to the job is wrong and it seems that BlueOcena is the cause.
I don't think anybody is actually using it but if there are some
objections, let me know.
Regards,
Davide
10:48 p.m.
I've removed the Blue Ocean plugin.
Let me know if there are still some problems.
I've also configured the jobs so that the website jobs can be executed
in parallel with the others.
This should make publishing of a page faster.
Cheers,
Davide
On Wed, Feb 22, 2017 at 10:47 AM, Davide D'Alto <davide(a)hibernate.org> wrote:
Hi.
I'm going to remove the BlueOcean plugin from Ci.
Currently, when we receive an email because a build is failed, the
link to the job is wrong and it seems that BlueOcena is the cause.
I don't think anybody is actually using it but if there are some
objections, let me know.
Regards,
Davide
11:24 a.m.
Thanks Davide!
All, we need your help now: please make sure that any job you enable
which uses any network port (like using Byteman, Arquillian, WildFly,
Databases, Elasticsearch, etc..) has set a "Job Weight" of 3.
When failing to do so, we'll see sporadic failures of both your job
and some other random job, so let's be nice to each other ;)
I think the setup is not perfect yet, as I've reviewed some failure
reports and they seem caused by an already running WildFly instance on
the same machine. So either we have a bad job configuration somewhere,
or a WildFly instance which didn't shut down correctly.
Thanks,
Sanne
On 22 February 2017 at 21:48, Davide D'Alto <davide(a)hibernate.org> wrote:
I've removed the Blue Ocean plugin.
Let me know if there are still some problems.
I've also configured the jobs so that the website jobs can be executed
in parallel with the others.
This should make publishing of a page faster.
Cheers,
Davide
On Wed, Feb 22, 2017 at 10:47 AM, Davide D'Alto <davide(a)hibernate.org> wrote:
> Hi.
> I'm going to remove the BlueOcean plugin from Ci.
>
> Currently, when we receive an email because a build is failed, the
> link to the job is wrong and it seems that BlueOcena is the cause.
>
> I don't think anybody is actually using it but if there are some
> objections, let me know.
>
> Regards,
> Davide
_______________________________________________
hibernate-dev mailing list
hibernate-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev
11:35 a.m.
I suspect the WildFLy instance was still active because I've aborted
the job or restarted Jenkins at the wrong time while
appying the changes.
Speaking about the job weight, I didn't change the releases job, I'll
let the owners of each job to do that but I think it would make sense
to cahnge their weight to 2 (unless one don't mind to wait).
Thanks,
Davide
On Thu, Feb 23, 2017 at 10:24 AM, Sanne Grinovero <sanne(a)hibernate.org> wrote:
Thanks Davide!
All, we need your help now: please make sure that any job you enable
which uses any network port (like using Byteman, Arquillian, WildFly,
Databases, Elasticsearch, etc..) has set a "Job Weight" of 3.
When failing to do so, we'll see sporadic failures of both your job
and some other random job, so let's be nice to each other ;)
I think the setup is not perfect yet, as I've reviewed some failure
reports and they seem caused by an already running WildFly instance on
the same machine. So either we have a bad job configuration somewhere,
or a WildFly instance which didn't shut down correctly.
Thanks,
Sanne
On 22 February 2017 at 21:48, Davide D'Alto <davide(a)hibernate.org> wrote:
> I've removed the Blue Ocean plugin.
>
> Let me know if there are still some problems.
>
> I've also configured the jobs so that the website jobs can be executed
> in parallel with the others.
>
> This should make publishing of a page faster.
>
> Cheers,
> Davide
>
>
> On Wed, Feb 22, 2017 at 10:47 AM, Davide D'Alto <davide(a)hibernate.org>
wrote:
>> Hi.
>> I'm going to remove the BlueOcean plugin from Ci.
>>
>> Currently, when we receive an email because a build is failed, the
>> link to the job is wrong and it seems that BlueOcena is the cause.
>>
>> I don't think anybody is actually using it but if there are some
>> objections, let me know.
>>
>> Regards,
>> Davide
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev
_______________________________________________
hibernate-dev mailing list
hibernate-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev
12:09 p.m.
On 23 February 2017 at 12:24, Sanne Grinovero <sanne(a)hibernate.org> wrote:
I think the setup is not perfect yet, as I've reviewed some
failure
reports and they seem caused by an already running WildFly instance on
the same machine. So either we have a bad job configuration somewhere,
or a WildFly instance which didn't shut down correctly.
There is a configuration problem with WildFly, partly due to an
arquillian bug, partly due to a bug in the wildfly-arquillian runner,
but both can be bypassed using this patch
http://pastebin.com/q2Nt0Ent
12:21 p.m.
Thanks Jordan!
Can you explain why we need to define the "managementPort" property?
Do you have a link to the issue maybe?
I'd like to understand that better, especially as we have several
other projects to patch.
Sanne
On 23 February 2017 at 11:09, Jordan Gigov <coladict(a)gmail.com> wrote:
On 23 February 2017 at 12:24, Sanne Grinovero
<sanne(a)hibernate.org> wrote:
> I think the setup is not perfect yet, as I've reviewed some failure
> reports and they seem caused by an already running WildFly instance on
> the same machine. So either we have a bad job configuration somewhere,
> or a WildFly instance which didn't shut down correctly.
There is a configuration problem with WildFly, partly due to an
arquillian bug, partly due to a bug in the wildfly-arquillian runner,
but both can be bypassed using this patch
http://pastebin.com/q2Nt0Ent
12:54 p.m.
First the comment needs to be outside the property, otherwise it's not
detected as a text-only node.
Secondly managementPort has to be set, because while arquillian passes
on jboss.socket.binding.port-offset but does not take it into account.
Because of port-offset the management port is moved from 9990 to
10127, but it's the runner framework doesn't know that. It gets stuck
waiting to confirm server start, but it's looking at port 9990 to
check it.
And it looks like I've made the patch with tabs, while the rest of the
file is spaces. Don't commit it mixed like that.
On 23 February 2017 at 13:21, Sanne Grinovero <sanne(a)hibernate.org> wrote:
Thanks Jordan!
Can you explain why we need to define the "managementPort" property?
Do you have a link to the issue maybe?
I'd like to understand that better, especially as we have several
other projects to patch.
Sanne
On 23 February 2017 at 11:09, Jordan Gigov <coladict(a)gmail.com> wrote:
> On 23 February 2017 at 12:24, Sanne Grinovero <sanne(a)hibernate.org> wrote:
>> I think the setup is not perfect yet, as I've reviewed some failure
>> reports and they seem caused by an already running WildFly instance on
>> the same machine. So either we have a bad job configuration somewhere,
>> or a WildFly instance which didn't shut down correctly.
>
> There is a configuration problem with WildFly, partly due to an
> arquillian bug, partly due to a bug in the wildfly-arquillian runner,
> but both can be bypassed using this patch
> http://pastebin.com/q2Nt0Ent
2612
days inactive
2613
days old
6 comments
3 participants
participants (3)
-
Davide D'Alto -
Jordan Gigov -
Sanne Grinovero