Re: [cbi-dev] CBI build farm grows by 40%!

["Daniel Megert" <daniel_megert@xxxxxxxxxx>]

>  Jiro

What does this
stand for?

Dani



From:
       "Mikaël
Barbero" <mikael.barbero@xxxxxxxxxxxxxxxxxxxxxx>
To:
       Common-build
Developers discussion <cbi-dev@xxxxxxxxxxx>
Date:
       02.04.2019
11:31
Subject:
       Re:
[cbi-dev] CBI build farm grows by 40%!
Sent
by:        cbi-dev-bounces@xxxxxxxxxxx

---

Hi Karsten, 

Xtext has been an early adopter for the
migration from the "old infra" https://ci.eclipse.org/xtext(the one with Jenkins OSS running on bare metal, with all the "noisy
neighbors" issues we know) to the new clustered infra (the one that
has been beefed up as Denis explained below). In that sense, you're a special
case. However, here is the global status and plan for the new infra. At
the end, I will comment on your specific Xtext case ;)

On this new clustered infra, we have
two kinds of Jenkins instances running: 

- CloudBees Jenkins (CJE) running at
https://jenkins.eclipse.org/
- and what we call Jiro (a custom distribution
of Jenkins OSS made by the Foundation Team) and tailored for running on
this cluster. Those instances also runs at https://ci.eclipse.org/(you can see the instances running with Jiro as they are labeled as "openshift"
instead of "hippX").

We're currently migrating all instances
from the "old infra" to Jiro. During the migration, projects
have 2 instances running: the old one on the old infra and a Jiro instance
on the new infra (available from the domain ci-staging.eclipse.org instead
of ci.eclipse.orgduring the migration to avoid conflicts with the old instance). Once the
migration is done and once the project validated that the Jiro instance
is ok, we archive the old instance and make the Jiro one available at ci.eclipse.org.

Note that all new instances (for new
or existing projects without a JIPP) have their JIPP created on the new
infra (i.e. they are Jiro JIPP).

We also have 49 instances running on
the new infra with CloudBees Jenkins (CJE). Those are mostly ee4j projects
and a couple of others (incl. Xtext). We don't plan to do a migration of
any ee4j project from CJE to Jiro yet. It's not a high priority because
they are already running on the new clustered infra. The priority is to
free up hippX machines to include them to cluster.

However, we will migrate the CJE early
adopters (like Xtext) who have 2 instances running both on the new infra
with CJE and on the old infra with Jenkins OSS. Those projects will be
migrated to Jiro. We will reach out to you when we'll tackle your project.
It will be a no-op for you for anything that currently runs on CJE. 

We are currently draining JIPPs on the
old infra on a machine by machine basis. This way we can decommission those
machines more quickly from the old infra and make them part of the build
cluster. 

Thanks.

Mikaël Barbero 
Team
Lead - Release Engineering |Eclipse
Foundation
📱
(+33) 642 028 039 | 🐦 @mikbarbero
Eclipse
Foundation:
The Platform for Open Innovation and Collaboration

Le 2 avr. 2019 à 10:15, Karsten Thoms
<karsten.thoms@xxxxxxxxx>
a écrit :

The Xtext project is still testing with
the CJE Jenkins instance https://jenkins.eclipse.org/xtext/, and we are finally able to build Xtext including tests there. Our main
build is at https://ci.eclipse.org/xtext/(hipp3). 

Does this mean that the project will
just stay at ci.eclipse.organd we should drop jenkins.eclipse.org/xtext?
One major effort there was to get the
UI tests running in a specific container. Can these Jenkinsfiles with Kubernetes
configuration be reused on Jiro then? (e.g. https://github.com/eclipse/xtext-xtend/blob/kth_issue1309_cbi/CBI.Jenkinsfile)

~Karsten


Am 02.04.2019 um 10:00 schrieb Olivier
Delcroix <odelcroi@xxxxxxxxx>:

Thanks !

On Tue, Apr 2, 2019 at 4:30 AM Denis
Roy <denis.roy@xxxxxxxxxxxxxxxxxxxxxx>
wrote:
Greetings,
I'm pleased to
announce that our OpenShift/Kubernetes build cluster now has 40% more CPU
power! We've added 96 CPUs to the cluster today.
Actually -- that's
more than 40% CPU power. We selected processors that are much, much faster
for single-threaded applications. In the past, we'd favour more CPU cores,
at the expense of lower clock speed.
As Fred and Mikaël
migrate Legacy JIPP instances to the Kubernetes-powered Jiro[1], we'll
soon be migrating the 10 massive Jenkins servers (known as hipp1 to hipp10)
to the OpenShift cluster.

We hope the new,
fast iron serves you well!

[1] https://github.com/eclipse-cbi/jiro


-- 
Denis
Roy
Director,
IT Services | Eclipse
Foundation, Inc.
Eclipse
Foundation:
The Platform for Open Innovation and Collaboration
Twitter:
@droy_eclipse

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

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

_______________________________________________
cbi-dev mailing list
cbi-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe
from this list, visit
https://www.eclipse.org/mailman/listinfo/cbi-dev
[attachment "signature.asc"
deleted by Daniel Megert/Zurich/IBM] _______________________________________________
cbi-dev mailing list
cbi-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe
from this list, visit
https://www.eclipse.org/mailman/listinfo/cbi-dev