Skip to main content

[Information] LUMI maintenance break 7-21 January, 2026

Thursday 22 January 16:00 CET (17:00 EET)
The LUMI web interface is back online.
Thursday 22 January 14:00 CET (15:00 EET)
System administrators are looking at the LUMI web interface issues. We’ll let you know as soon as it becomes available again.
Wednesday 21 January 21:10 CET (22:10 EET)
The LUMI web interface will remain unavailable until Thursday 22 January.
Wednesday 21 January 21:00 CET (22:00 EET)
LUMI is back into production.
Almost all software components on LUMI have received an update during the system maintenance. The LUMI User Support Team has prepared a page listing the different updates and their consequences here https://lumi-supercomputer.github.io/update. We will keep updating this page in the next few weeks. Please read this page carefully before restarting your work on LUMI and keep an eye on it the in the coming weeks.
A large percentage of GPU jobs submitted before the maintenance break are expected to fail due to changes in the software environment. Therefore, these jobs have been placed on hold to allow you time to inspect them before deciding to cancel them, or, if you wish, release them and monitor their progress. After a one-week grace period, any held jobs will be removed from the queues. Instructions can be found here https://lumi-supercomputer.github.io/LUMI-training-materials/User-Updates/Update-202601/#how-to-get-running-again. Some CPU-based jobs may also fail, but it is expected that this failure will be immediate, without wasting billing units, so these jobs will be released by the system administrators.
On Wednesday 28 January 2026 at 13:00 CET (14:00 EET) we’ll have the regular last Wednesday of the month LUMI user coffee break where we will answer burning questions about the update via Zoom https://cscfi.zoom.us/j/65727034273?pwd=VEdtY2trVUVKTEhxajZMbFhETWV2Zz09.  And on Wednesday 11 February 2026 at 13:00 CET (14:00 EET), we will organize a webinar where HPE and AMD staff will give a presentation about the main changes in the environment, again via Zoom https://cscfi.zoom.us/j/65727034273?pwd=VEdtY2trVUVKTEhxajZMbFhETWV2Zz09. For both events it will be possible to ask questions via a collaborative HedgeDoc using the link https://siili.rahtiapp.fi/lumi-coffee-break?both#.
If you experience issues, do not hesitate to contact the LUMI User Support via the web forms https://lumi-supercomputer.eu/user-support/need-help/.
LUMI-O will be available during the maintenance break except 19-21 January.
The LUMI system will be offline for maintenance starting on Wednesday, 7 January, 2026.
LUMI login nodes, compute nodes, and all the storage partitions won’t be accessible with the exception of the object storage partition LUMI-O which will remain available most of the maintenance period. A network maintenance will be done at the same time that can affect access to LUMI-O for a couple of days (we will communicate further as soon as these days are confirmed).
We expect the system to be back in production on Wednesday, 21 January, 2026.
Significant parts of the system software will be updated in order to get a more stable and up-to-date system after the maintenance break. It includes the base OS (upgraded to SUSE 15 SP6 and the equivalent COS version), and the Slingshot host software. ROCm will be updated to version 6.3 and the programming environment to version 25.03 which is based on ROCm 6.3 and, for the Cray compilers, clang/LLVM 19.
The driver for ROCm 6.3 officially supports ROCm versions from 6.1 onward, but already benefits from the attempts of AMD to offer a longer support window for ROCm userland libraries on ROCm drivers so should be able to support ROCm versions up to 7.0.
As part of the update, older versions of the Programming Environment that are not fully functional anymore will be removed from the system. However, 23.09 (CCE 16), and 24.03 (CCE 17) will remain available for now.
Please, don’t hesitate to contact us if you need any further information https://lumi-supercomputer.eu/user-support/need-help/