Messages & Announcements

  • 2013-10-16:  Sandhills changes (some just for now)
    Category:  General Announcement

    Sandhills was recently hacked via what appears to have been a local exploit (i.e. someone's account info was stolen). As a result, several changes have been implemented; you may not have noticed a new login node has been deployed. Forensics continue, but the attack appears to have been contained with limited impact. We have updated and changed several things on the login node. Users are asked to test their accounts and applications. Anything requiring connections to another machine may need some further maintenance -- please let hcc-support@unl.edu know asap if you need assistance or notice anything unusual.

    Further updates will be sent out as needed.

    This does not affect tusker.


  • 2013-10-10:  SANDHILLS is available for use.
    Category:  General Announcement

    Sandhills is available for use after scheduled maintenance. The nature of the maintenance required all running jobs to be halted. If you have running or waiting jobs you will need to resubmit them. -tharvill


    Sandhills is available for use after scheduled maintenance. The nature of the maintenance required all running jobs to be halted. If you have running or waiting jobs you will need to resubmit them. -tharvill

  • 2013-09-26:  Tusker Downtime Update
    Category:  General Announcement

    During today's planned Tusker outage there were some unexpected problems during the slurm upgrade.

    All precautions were observed in an attempt to preserve jobs in the queue, however the job state file created by the prior release version of slurm was not recoverable by this most recent slurm release. As a result, all jobs in the queue have been lost and will need to be resubmitted.

    Please accept our sincere apologies as we understand the inconvenience this unexpected software failure causes those users with jobs in the queue.


    During today's planned Tusker outage there were some unexpected problems during the slurm upgrade.

    All precautions were observed in an attempt to preserve jobs in the queue, however the job state file created by the prior release version of slurm was not recoverable by this most recent slurm release. As a result, all jobs in the queue have been lost and will need to be resubmitted.

    Please accept our sincere apologies as we understand the inconvenience this unexpected software failure causes those users with jobs in the queue.

  • 2013-09-26:  SANDHILLS downtime announcement (maintenance)
    Category:  Maintenance

    SANDHILLS will be shutdown sometime after 11:00am on October 10, 2013 in order to upgrade infiniband device drivers. All running jobs will be halted. SANDHILLS should be available for use again no later than the morning of October 11. -tharvill


    SANDHILLS will be shutdown sometime after 11:00am on October 10, 2013 in order to upgrade infiniband device drivers. All running jobs will be halted. SANDHILLS should be available for use again no later than the morning of October 11. -tharvill

  • 2013-09-19:  Tusker downtime September 26 for maintenance
    Category:  Maintenance

    Jobs which cannot complete before September 26 at 10:00am will be held in queue until the maintenance is complete. We anticipate this work will be finished by 6:00pm the same day.

    If this timing significantly disrupts time-critical work (conference deadline, class, etc.) please contact David Swanson, Director HCC, at 472-5006.

    Details:

    New network gear will be brought online at the PKI data center on September 26th. This gear will replace the network core that has been servicing the Firefly and Tusker clusters.

    To minimize impact to running jobs, we are declaring a downtime for Tusker to complete this work. We will use this maintenance window to update software components across the Tusker cluster. The Tusker login node will also be updated and will require users to log off. Users will be denied access to the Tusker login node until the maintenance is completed.

    Network connectivity to the Firefly login node will also be impacted throughout the maintenance window.


    Jobs which cannot complete before September 26 at 10:00am will be held in queue until the maintenance is complete. We anticipate this work will be finished by 6:00pm the same day.

    If this timing significantly disrupts time-critical work (conference deadline, class, etc.) please contact David Swanson, Director HCC, at 472-5006.

    Details:

    New network gear will be brought online at the PKI data center on September 26th. This gear will replace the network core that has been servicing the Firefly and Tusker clusters.

    To minimize impact to running jobs, we are declaring a downtime for Tusker to complete this work. We will use this maintenance window to update software components across the Tusker cluster. The Tusker login node will also be updated and will require users to log off. Users will be denied access to the Tusker login node until the maintenance is completed.

    Network connectivity to the Firefly login node will also be impacted throughout the maintenance window.