Dear Cyclone and Grunch Users,

We plan to have maintenance and downtime on Local HPC next week, from Monday to Wednesday 10-12 July, 2023. We will need full downtime with all the servers in the rack turned off. We will then restart them several times for doing firmware upgrades.
Cyclone and Grunch servers will be affected by this.
Please save your work before Monday.
Don't hesitate to contact us if you have any questions via hjelp.uib.no.

[Update on 17.07.2023] Most of the Cyclone functionalities are now restored after the maintenance, but there is a chance that some of you may experience issues following the introduced changes. We anticipate that we will need to remount and restart the storage systems in the coming days since some users have already reported missing file access or similar problems.

On Tuesday 18th of July we expect visit from Dell Field Engineer, who will work with service of Lustre storage (/shared and /home folders).We plan to restart the storage servers on that day. Planned work is important for integrity of the storage.

Dear Cyclone and Grunch Users,

Datacenter is planning a downtime during the weekend for continuity tests.

cyclone and grunch server will be taken down and will not be accessible during this time.

We are sorry for the inconvenience.

Both the servers are expected to be online on Monday.

Best Regards

Scientific Computing Group

Update on 24.04.2023: Cyclone and Grunch are up and running.

We are currently experiencing issues with writing to files under the /shared directory, which includes users' home folders on Cyclone and Grunch. We are looking into the issue, and apologize for the inconvenience.

Edited Dec 1: We still haven't been able to resolve the issues, and will perform a reboot of the systems at 11:00 today (Dec 1).

Edited Dec 2: It seems that we were able to resolve the issues. Please let us know if any of the systems do not perform as expected.

Dear Cyclone and Grunch Users,


cyclone.hpc.uib.no and grunch.hpc.uib.no are currently unavailable due to file system issues.  Our system operators shall resolve the issues at the earliest.  

More updates shall follow during the course of the day. 

Feel free to contact us via hjelp.uib.no if you have any questions.

Apologies for the inconvenience.

Edited: The issue on Cyclone is resolved now and the system is up and running. 
grunch.hpc.uib.no  is still unaccessible and  we shall update the status tomorrow.

Edited Jul 7: The issue on Grunch is now resolved, and both systems are up and running again.

Best regards,

The IT department and the Scientific Computing Group

 

Dear Users of Cyclone, 

cyclone.hpc.uib.no is  inaccessible at the moment. We need to restart the system.

More information will follow during the course of today.

EDITED:  The issue is resolved. You can use the system again. Please, let us know if it looks like there is still something wrong with cyclone.

Best regards,

The IT department and the Scientific Computing Group

Dear users,

We were made aware that to cyclone and leo could not be accessed via ssh from the UiB network last weekend. The machines were restarted yesterday, and now seem to respond as expected. Please contact us through hjelp.uib.no if you experience any further issues.

We are sorry for the inconvenience.

Best regards,
The Scientific Computing group at the IT department

Dear Grunch users,

We plan to have maintenance and downtime of grunch.hpc.uib.no for about an hour this Monday. Even as they share some of the same filesystems, only grunch will be restarted, and cyclone will not be affected.

Edit 2022-04-11 15:20 : We cancelled the maintenance and it won't happen anymore this or next week. When it will, we'll put up a new notice. Sorry for the false alarm.


Best regards,

The IT department and the Scientific Computing Group

Dear users of cyclone,

cyclone.hpc.uib.no is inaccessible at the moment. We will probably need to restart the system.

More information will follow during the course of today.

edit 2021-09-15 13:00: Because of unresponsiveness of the system we needed to restart cyclone.hpc. We do not know the cause yet and are still looking into it. You can use the system again. Please, let us know if it looks like there is still something wrong with cyclone.

Best regards,

The IT department and the Scientific Computing Group

Dear users of cyclone,

We will have an emergency downtime until further notice because of issues with the Lustre filesystem on cyclone.hpc.uib.no. Please, save any work and log out.

Feel free to contact us via hjelp.uib.no if you have any questions.

Apologies for the inconvenience.

Update 18.06: cyclone is up and running. Yesterday we thought that there might be something wrong with the local Lustre filesystem, which includes /home and /Data/gfi/stormrisk. To reduce the risk of data loss, we brought the systems offline. However, everything appears to be working reliably after all.
In addition, the backup of /home does not always finish. We are still looking into this problem, but you may login and use cyclone again.

Best regards,

The IT department and the Scientific Computing Group