Updates

Event Date	Summary
	June 30 @ 5:36PM PST - We think we have sorted out the cooling issue. We are currently in the process of running some final cooling tests and expect to begin the process of bringing up Cedar nodes at 7PM PST. We thank you for your continued patience!
	June 30 @ 9:37AM: SFU Facilities Services is on-site and performing maintenance on the cooling system to try and determine why that unit are intermittently power cycling the fans and pumps. HPC services remain offline while we sort through the issue and will provide updates as they are available.
	Cedar is now Available With Conditions Globus services are now back online. The SFU datacenter continues to experience cooling issues. While temperatures have remained stable over the weekend, we are keeping all HPC services offline, with the exception of Cedar Cloud, dCache, and Globus services. Access to login nodes is now available. Important information Please note that all of the file systems are now using the new storage infrastructure. If users encounter any issues, they should not hesitate to submit a support ticket. However, here is some important information...Project upgrade: /project filesystem data was migrated to our new storage, if you find any missing data, or other concerns, such as quota, support is available to assist. Benefit of this migration is the new Project Quota, rather than SGID of project; this means: all of the files inside the parent project folder belong to the quota. This eliminates issues often faced with tools, such as: git, mv, or tar Possible reasons for out-of-quota issues: The new project does not have file compression enabled, this is to be expected. /project now uses Project Quota, so all files inside of the project apply against the PI's quota, whether they belong to the UID/GID of the PI or their users. This will most likely affect projects that had files from old projects that they had moved over or left behind. Thank you for your patience and understanding

Incident description

Service	Incident status	Start Date	End Date
Cedar	Closed

Created by Andre Poley on

Title

Planned Outage - Arrêt planifié

Summary

Update June 29, 2025: Globus services are now back online. The SFU datacenter continues to experience cooling issues. While temperatures have remained stable over the weekend, we are keeping all HPC services offline, with the exception of Cedar Cloud, dCache, and Globus services.
Please note that all of the file systems accessible via Globus now use the new storage infrastructure. If you encounter any issues, please do not hesitate to submit a support ticket.
Access to login nodes is now available.
Thank you for your patience and understanding.

UPDATE June 28, 2025: SFU data centre is currently experiencing cooling issues. The temperature rose significantly over the past 24 hours. We are keeping all HPC services offline except Cedar cloud, dCache services and Globus to access old /project filesystem (readonly).
UPDATE June 24, 2025: Filesystem (/home, /scratch) is being updated by the vendor and is no longer available until Friday, June 27. /project will still be available for readonly.
UPDATE: /scratch and /home are now accessible in read only mode on globus
Note: If you need urgent read-only access to the project filesystem you can use globus to access your files. You can find more information about globus in the wiki here: https://docs.alliancecan.ca/wiki/Globus#Using_Globus
go to https://globus.alliancecan.ca
login and consent to dialogues
go to File Manager
select the Collection 'computecanada#cedar-globus'
provide a full path to your file using either the project name or the numeric id of the project you want to access. You can get the numeric id by running the command 'id' on any alliance login node (if you dont specify a path you might get an error message)

Cedar will be unavailable from 2025-06-16 to 2025-06-30. During this time, you will not be able to log in or run jobs on the cluster. Any jobs that are running at the time of the outage will be terminated and you will need to re-submit them once the cluster is back online.

======

Cedar ne sera pas disponible 2025-06-16 de 2025-06-30. Durant cette période, les utilisateurs ne pourront pas se connecter ni exécuter des tâches sur Cedar. Toute tâche en cours au moment de l’arrêt de service sera arrêtée et devra être soumise à nouveau une fois les travaux terminés.

Updates

Incident description

Created by Andre Poley on document.scripts[document.scripts.length - 1].parentNode.appendChild(document.createTextNode(change_date_full("2025-06-13 15:45", "en")));

Title

Summary

Updated by James Peltier on document.scripts[document.scripts.length - 1].parentNode.appendChild(document.createTextNode(change_date_full("2025-07-01 05:53", "en")));

Created by Andre Poley on

Updated by James Peltier on