Messages & Announcements

  • 2013-02-13:  FIrefly Retirement this summer
    Category:  General Announcement

    Dear Colleagues,

    The Firefly cluster will be retired in the next few months. Please move workflows to Tusker as soon as possible. Data on /work is not backed up -- only expendable data should reside there. If you need assistance with this move, please contact hcc-support@unl.edu.

    Tusker has been in place for the better part of a year, and a second hardware acquisition is planned for later this summer. The remainder of Firefly will be retired when the new hardware installation date nears. This is currently planned for June, 2013.

    Further, a significant hardware failure, while not hoped for or expected, would likely result in immediate retirement of Firefly. Dell no longer produces SC1435 nodes. Panasas no longer sells the storage product we continue to use. Force10's switch fabric is out of warranty, and Cisco no longer produces Infiniband of any kind. This email is to emphasize the need for current Firefly users to plan for this upcoming event. Expect further communication as the retirement date approaches.

    If you have any questions or particular concerns, please let us know. In the mean time, continue to use Firefly, but caveat emptor -- its days are numbered.

    Best regards,
    David

    David R. Swanson, Director
    Holland Computing Center
    University of Nebraska




  • 2013-01-04:  Sandhills login node network outage, service restored
    Category:  System Failure

    Between 5:00pm and 7:00pm today, there was a network outage affecting access to the Sandhills login node. Network service has been restored. Running jobs were not affected.


    Between 5:00pm and 7:00pm today, there was a network outage affecting access to the Sandhills login node. Network service has been restored. Running jobs were not affected.

  • 2012-12-17:  TUSKER: service restored
    Category:  General Announcement

    Tusker maintenance complete, the system is now available for use.


    Tusker maintenance complete, the system is now available for use.

  • 2012-12-09:  Tusker NFS restored, outage affected /home and /util
    Category:  System Failure

    The NFS server for Tusker stopped responding Sunday morning. This caused access to /home and /util to stall, as well as affecting SSH logins. After a reboot, the NFS service appears to be working normally. We apologize for the inconvenience.


    The NFS server for Tusker stopped responding Sunday morning. This caused access to /home and /util to stall, as well as affecting SSH logins. After a reboot, the NFS service appears to be working normally. We apologize for the inconvenience.

  • 2012-11-26:  Tusker Shutdown Dec. 17
    Category:  Maintenance

    A December 17th shutdown is planned for Tusker. During the shutdown we will be performing maintenance on, and upgrading, the Lustre storage system which supplies the /work filesystem. We request users to remove as much data as possible from the /work filesystem between now and then to 1) ensure it will be possible to wait until after finals week (Dec. 17) and 2) to help minimize the time required to perform the upgrade. Additional disks will be added to the storage array to improve reliability of the system -- the less data there is to move and/or re-stripe, the faster this process will be.

    Tusker will be made available immediately after this maintenance is complete.


    Four new disk enclosures will be added to the Lustre file system serving Tusker. Details are being worked out in discussions with the vendor (Terascala), but the end result will be a more robust, and likely slightly larger, filesystem. Right now multiple RAID 5 disk sets are in service (as was required by the vendor). HCC has succeeded in negotiating a change to RAID 6, but the migration path is still being finalized.

Pages