There is a new featured template for Ubuntu 20.04 LTS amd64 (ubuntu-20.04-lts). This template subsumes the former template Ubuntu 20.04 amd64, which is no longer featured.

Changes with respect to the former template:

• the initialization service for ‘~ci/.ssh/authorized_keys` is fixed, and the contents of this file will no longer be reset on each boot;
• cloud-init is now configured to consider CloudStack provided user-data by default;
• packages are upgraded to their last available version.

Thank you very much for using CI.

CloudStack is now updated to 4.16. You should now be able to start your virtual machines again via the portal ci.inria.fr or via CloudStack API or web interface.

You may refer to the portal documentation: https://inria-ci.gitlabpages.inria.fr/doc/page/web_portal_tutorial/#slaves

Thanks to this update, Mac OS Catalina and Mojave are supported, and nested KVM is enabled.

Thank you very much for using CI.

CloudStack is now updated to 4.16. You should now be able to start your virtual machines again via the portal ci.inria.fr.

You may refer to the portal documentation: https://inria-ci.gitlabpages.inria.fr/doc/page/web_portal_tutorial/#slaves

We still have a problem that prevents users to log in to CloudStack by the web interface. Only the CloudStack web interface seems to be affected: you are still able to manage your virtual machines via the portal or the CloudStack API. We will fix the connection to the web interface as soon as possible.

Thank you and sorry for the inconvenience.

All VMs hosted on ci.inria.fr will be interrupted on Tuesday afternoon 2023-02-07, 2pm-6pm, for CloudStack maintenance. CloudStack will be upgraded to 4.16, and nested KVM will be enabled. The actual interruption may be shorter, we will publish an announce when the platform will be back. You will have to restart the VMs manually via ci.inria.fr or CloudStack. Thank you!

The shared storage for Jenkins instance is back to normal since last Thursday. If your Jenkins instance is down due to disk space outage, restarting the instance should now work again. Sorry for the inconvenience, we will try in the next few days to take counter-measures to prevent these kinds of problems in the future.

The shared storage for Jenkins instance is full: this probably prevents all Jenkins instances to work properly. We are currently investigating why we have difficulties to free space and I hope we will be able to restore the service very soon!

The certificate of the LDAP server was expired this morning and has been renewed. You may have had some difficulties to connect to ci.inria.fr, CloudStack and Jenkins instances this morning but everything should work again now! Sorry for the inconvenience.

You may have received this morning some emails inviting you to extend the expiration date of your CI virtual machines. As you may have noticed, the URLs were wrong: /prolongation was missing at the end. You may either use the fixed URL, or the “Extends” button in the slave table in the web interface of your CI project.

Sorry for this inconvenience, and thank you for extending your VMs!

We are pleased to announce that shared docker runners are now available for gitlab.inria.fr!

Thanks to everyone who helped testing the service during the (very long!) beta period. You can right now enable the shared runners in the CI/CD settings of your Gitlab projects.

See the shared runners documentation for more details. The three following docker runners are available:

• small (1 CPU, 2048 MB RAM), tags: ci.inria.fr, linux, small
• medium (2 CPUs, 4096 MB RAM), tags: ci.inria.fr, linux, medium
• large (4 CPUs, 8096 MB RAM), tags: ci.inria.fr, linux, large

The runners will handle jobs declared in the .gitlab-ci.yml file of the project repository, at the condition that the jobs have at least one tag among ci.inria.fr, linux and either small, medium or large (these three last tags are exclusive: they cannot be used together for the same job since no runner will match more than one of them).

We would like to send a big thank you to Anthony Baire for all his efforts in developping this service, and we would like to thank also the CI infrastructure team of the Inria DSI who helped in deploying the platform.

Thank you very much for using ci.inria.fr, and happy hacking!

For a few days and for an unknown reason, Mac OS X VMs are continuously rebooting when they are newly created or when their offer is changed. We are aware of the problem and are trying to fix it as soon as possible.

Thank you for your understanding!

Each VM hosted by ci.inria.fr has now an expiration date, fixed in one year from now for current VMs, and one year after the creation date for future VMs. This rule will make us able to clean unused VMs and save ressources for new projects.

These expiration dates are freely renewable, using the dedicated button “Extend” in the VM list associated to a project in the portal ci.inria.fr.

When VMs are about to expire, project admins will be reminded three times by email (with a link to extend the expiration date) starting three weeks before the expiration. Expired VMs will be deleted and project admins will be notified (there will be a 24h delay before expunging, during which the VM can be recovered).

Thank you for your understanding!