[Issues] Peregrine maintenance extended

Ewout M. Helmich e.m.helmich at rug.nl
Thu Nov 8 11:52:16 CET 2018

Hi everyone,

Please note that there are issues with the planned Peregrine 
maintenance. It will be extended at least to tomorrow morning, see below.




Dear Peregrine users,

The process of copying /data from the Data Handling storage to the new 
Peregrine storage is going much slower than anticipated, due to a few 
crashes of the copying process. This forces us to restart the 
synchronization process from the beginning, and this will take some time.
Because of the way this process works, we have decided not to allow 
access to the cluster, since doing so can result not only in jobs 
crashing (because of missing files), but also in file conflicts if new 
files are created on the Peregrine storage with the same names as files 
that still need to be copied over from the Data Handling storage.

We understand that the prolonged downtime of the cluster is adversely 
impacting your work, but, on the whole, it is the least of two evils. We 
apologize once again for this, and hope that you understand our reasoning.

As regards an estimate of when Peregrine will be back online, the 
earliest time would be tomorrow morning, if everything works as it 
should. We are also preparing for a worse case scenario. We will have 
more information for you tomorrow.

Best regards,

Fokke Dijkstra
The HPC Team



More information about the Issues mailing list