The file system is now back online.
Some of the compute nodes may have lost connectivity to /scratch
for a few minutes, so if your jobs were actively using data in $SCRATCH
or $GROUP_SCRATCH
, please check the output of your jobs for I/O errors, and resubmit the ones that may have failed. We apologize for the trouble this may have caused.
We’re investigating an issue on /scratch
. The file system is currently unavailable.