**ALL FILES IN /SCRATCH WILL BE DESTROYED ON January 9, 2018**
We will be performing the storage cut-over on Tuesday, January 9 (starting at 12:01 a.m.), ahead of schedule. We were able to complete a complete first copy of the storage to the new filesystem. Since we are ready well ahead of schedule, we are making the compromise of making the switch over on shorter notice in order to get it done before the beginning of the semester and the load on the cluster increases. The cluster utilization has been lighter over the break and this has allowed us to copy all your data faster. By doing it sooner, we should have a smaller outage (so we do not have to catch up with all the work done). We are also seeing increasing failures on our test environment and need to retire it ASAP.
***IMPORTANT*** The /scratch filesystem will not be copied and all the files in /scratch WILL BE DESTROYED when we decommission the old filesystem and shred all the hard drives.
The quota system will also change and we will give specific information during the switchover.
Users of the /storage/HPC will need to migrate their files after the outage. We will be contacting groups individually for this.
We have 11 more GPU’s that were on back-order on site and will be adding them soon.
On a lighter note, 2017 saw the following numbers:
* 4 PB of data transferred in and out of the Research Network
* 25 PB of network traffic moved across the cluster switches
* 3.5 million jobs completed
* 32 million core-hours of scientific computation performed with over 75% core utilization over the entire year (and growing)
* Over 700 with on average 130 users active each month.
Thanks once again for your patience and understanding as we work to improve our services.
— Jan. 3, 2018