Status of System

Lise at NHR@ZIB

timebar.png

File system replacement

15.01. 16:00 - Data migration phase has been extended. On Thursday Jan. 23rd 10:00am, we will start a maintenance on system Lise for several hours. The old WORK storage will be deleted. Until Jan. 22nd the old WORK directories are available under /old-scratch/usr, /old-scratch/projects and /old-scratch/share.

More information you find on File system replacement.

More trouble with the new file system - fixed

15.01. 16:00 - System Lise is back in production. Please remember the maintenance on Jan 23rd.
14.01. 18:00 - Problem found: broken gateway server between our OPA and Infiniband network caused cascade, 1st, of disabled GPFS clients (compute nodes) and, 2nd, of disconnected storage servers.
14.01. 10:00 - Problems still persist. Data-migration deadline extended (see below) due to unreliable file system access.
13.01. 16:00 - /home and /scratch still can be accessed (most of the time). In contrast, /sw is not accessible any longer. Hence, module files cannot be loaded.
13.01. 10:00 - Slurm/job submission is closed now. Login is open for migration but data access may be interrupted sometimes. An IBM team is debugging.
10.01. 00:15 (initial)- We are experiencing GPFS problems (like before Christmas). For now, we leave most nodes open while investigating. IBM (vendor providing file system support) ticket is issued.

Related pages