Smalltalk › Pharo › Pharo Smalltalk Developers

Re: [ci-announces] IMPORTANT: Virtual Machines loss

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

3 messages Options

Volkert

Re: [ci-announces] IMPORTANT: Virtual Machines loss

Out of curiosity: What is the impact on the Pharo Project? Is the build process is interrupted?

Volkert

On 28.07.2017 19:18, Christophe Demarey wrote:

Dear CI users,

Due to a manipulation error, a lot of Virtual Machines hosted on CloudStack were destroyed and are not recoverable!
For the projects being impacted, there is no other solution than recreate new VMs. Please find as attachment the list of virtual machines that are not affected.

Jenkins servers are not concerned by this issue, so all your jobs and history are safe.

We are deeply sorry for this failure.
Everything was done in an effort to retrieve the lost VMs, but it has proven to be an impossible task and the VMs cannot be restored. We are now working on setting up safeguards to prevent such situations.

With our sincere apologies,
The CI team.

Guillermo Polito

Re: [ci-announces] IMPORTANT: Virtual Machines loss

I don't want to be pesimistic, but this means we don't have CI until we restore all slaves. From the Pharo point of view, we lost almost all but 1 or 2 slaves in different ci servers. Check:

- https://ci.inria.fr/pharo/

- https://ci.inria.fr/pharo-contribution/

- https://ci.inria.fr/pharo-ci-jenkins2/

I'd like to make for real a call for help in here. We have, for each project (pharo, pharo-contribution, ci-jenkins2), to set-up slaves for windows linux and mac. This means we have to create slaves, check they can connect to jenkins (and otherwise configure them, install dependencies (windows through remote desktop...)).

I propose to do a meeting on discord #development tomorrow, monday, at 14h CEST (UTC/GMT+2, Paris/Brussels time) to organize a task force. Like that we can split tasks and share knowledge on this to have less bottlenecks, what do you think?

From my side, I'll try to spend monday morning seeing how we can automatize most of these things...

Guille

On Sun, Jul 30, 2017 at 8:11 AM, Volkert <[hidden email]> wrote:

Out of curiosity: What is the impact on the Pharo Project? Is the build process is interrupted?

Volkert

On 28.07.2017 19:18, Christophe Demarey wrote:

Dear CI users,

Due to a manipulation error, a lot of Virtual Machines hosted on CloudStack were destroyed and are not recoverable!
For the projects being impacted, there is no other solution than recreate new VMs. Please find as attachment the list of virtual machines that are not affected.

Jenkins servers are not concerned by this issue, so all your jobs and history are safe.

We are deeply sorry for this failure.
Everything was done in an effort to retrieve the lost VMs, but it has proven to be an impossible task and the VMs cannot be restored. We are now working on setting up safeguards to prevent such situations.

With our sincere apologies,
The CI team.

Guille Polito

Research Engineer

French National Center for Scientific Research - http://www.cnrs.fr

Web: http://guillep.github.io

Phone: +33 06 52 70 66 13

demarey

Re: [ci-announces] IMPORTANT: Virtual Machines loss

Le 30 juil. 2017 à 10:39, Guillermo Polito <[hidden email]> a écrit :

I don't want to be pesimistic, but this means we don't have CI until we restore all slaves. From the Pharo point of view, we lost almost all but 1 or 2 slaves in different ci servers. Check:

- https://ci.inria.fr/pharo/
- https://ci.inria.fr/pharo-contribution/
- https://ci.inria.fr/pharo-ci-jenkins2/

I'd like to make for real a call for help in here. We have, for each project (pharo, pharo-contribution, ci-jenkins2), to set-up slaves for windows linux and mac. This means we have to create slaves, check they can connect to jenkins (and otherwise configure them, install dependencies (windows through remote desktop...)).

I will create most of slaves today for all projects.