Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [mdt-papyrus.dev] Additional Job Queue on Hudson

Hi,

 

An up to date example of the current problem see [1]

 

The 5 min sysml-gerrit job will have to wait more than 1 hour.

Adding a dedicated runner for all jobs that take less than 15 min would really help.

 

Regards,

Benoit

1 :

 

De : mdt-papyrus.dev-bounces@xxxxxxxxxxx [mailto:mdt-papyrus.dev-bounces@xxxxxxxxxxx] De la part de LETAVERNIER Camille
Envoyé : mercredi 21 octobre 2015 10:16
À : Papyrus Project list
Objet : [PROVENANCE INTERNET] Re: [mdt-papyrus.dev] Additional Job Queue on Hudson

 

Hi,

 

Only webmasters can help here :)

 

Personnally I’d even go for a specific HIPP instance for Gerrit builds, and keep all other builds on Papyrus HIPP. We already increased the number of executors from 2 to 4, which helped at the time, but I don’t know how far we can go.

 

Camille

 

De : mdt-papyrus.dev-bounces@xxxxxxxxxxx [mailto:mdt-papyrus.dev-bounces@xxxxxxxxxxx] De la part de Ed Willink
Envoyé : mercredi 21 octobre 2015 10:09
À : mdt-papyrus.dev@xxxxxxxxxxx
Objet : Re: [mdt-papyrus.dev] Additional Job Queue on Hudson

 

Hi

The HIPPs were originally created because Hudson or its plugins have concurrency bugs. e.g certain activities share a global /tmp filespace so that a file owned by one user's job can block another user's job. This happened quite regularly, more than weekly for me. So much so that a failure on Hudson didn't count until it was repeated. On the HIPPs the situation is drastically improved. I think I have only one unexplained failure in over a year; probably a result of concurrent P2 repo evolution.

If you need two jobs, I suggest that perhaps two HIPPs might me better.

    Regards

        Ed Willink

On 21/10/2015 08:26, LE FEVRE FRANCOIS wrote:

Dear all,

 

As you have noticed we have create 2 additional components SysML and RobotML: specific bugzilla category, specific repository.

We would like to know if it is possible also to create a dedicated job queue for each of them.

Why?

For instance, at this stage, developments on SysML are blocked due to the unavailability of job queue of Papyrus.

Each Papyrus gerrit job last at least 55 minutes, whereas the gerrit jobs on SysML last 5 minutes.

 

If no, why?

If yes, what is the processus to create the 2 job queues and to configure them to take job only from the SysML and RobotML component/repository respectively.

 

Thanks for your help.

 

cid:part1.06060709.08040600@ericsson.com

Francois LE FEVRE

Research Engineer

Commission for Atomic Energy and Alternative Energies (CEA)

Model-driven Engineering for Embedded Systems Laboratory (LISE)

Projet: Papyrus: https://www.eclipse.org/papyrus/

cid:part5.02030405.05090707@ericsson.com

Commissariat à l’énergie atomique et aux énergies alternatives (CEA)

Paris-Saclay Campus - Nano-INNOV | Bât. 862-PC1087 | F-91191 Gif-sur-Yvette Cedex

T. +33 (0)1 69 08 49 86  |  F. +33 (0)1 69 08 83 95  |

francois.le-fevre@xxxxxx   |  Blog: http://biocamp.blogspot.fr/

 

 



_______________________________________________
mdt-papyrus.dev mailing list
mdt-papyrus.dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/mdt-papyrus.dev



-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.6172 / Virus Database: 4450/10860 - Release Date: 10/20/15

 


Back to the top