Messages & Announcements

  • 2016-12-15:  Sandhills shutdown Dec. 21 - Jan. 4
    Category:  Maintenance

    Campus chilled water will be shut down completely from Dec. 21 to Jan. 4. Since the Schorr machine room relies heavily upon this supply for cooling, and since this overlaps with the holiday shutdown, we plan to shut down most of Sandhills over this time period. We will leave up the filesystems and login node, but all compute (worker) nodes will be turned off to avoid damage to any machines. If you have any questions or concerns please contact hcc-support@unl.edu.


    HCC has no control over this chilled water outage, and will take the above action since it will result in a significantly lower cooling capability.

  • 2016-10-25:  HCC: debugging tools disabled
    Category:  General Announcement

    Debugging tools (gdb, strace, etc) have been temporarily disabled on HCC clusters until further notice. We will re-enable the debugging subsystem once the OS vendor releases a fix for a memory bug in the OS kernel. A follow-up announcement will be made at that time.


    Debugging tools (gdb, strace, etc) have been temporarily disabled on HCC clusters until further notice. We will re-enable the debugging subsystem once the OS vendor releases a fix for a memory bug in the OS kernel. A follow-up announcement will be made at that time.

  • 2016-10-13:  Recovery: Anvil Unexpected Downtime Finished
    Category:  General Announcement

    We just restarted the Ceph OSDs and theyâre back on line now. The Ceph filesystem is available now. The performance may still be degraded until the recovery finished. Please let us know if you have any problem.


    We just restarted the Ceph OSDs and theyâre back on line now. The Ceph filesystem is available now. The performance may still be degraded until the recovery finished. Please let us know if you have any problem.

  • 2016-10-13:  Anvil unexpected downtime
    Category:  System Failure

    We are currently experiencing a failure in our Ceph filesystem and are working to get it back up as soon as possible. The VMs on Anvil may be slow or unresponsive.


    We are currently experiencing a failure in our Ceph filesystem and are working to get it back up as soon as possible. The VMs on Anvil may be slow or unresponsive.

  • 2016-10-12:  File purge warning, all HCC resources
    Category:  General Announcement

    This notice affects users with data on the /work filesystem of all HCC resources (Crane, Sandhills and Tusker). Files not accessed for over 6 months will be removed this week. This will now be enforced as an ongoing policy. This was originally announced for all HCC resources September 29, and no more warnings will be sent unless the policy or situation changes. These files will not be recoverable. All of /work remains subject to further file removal if conditions so demand (i.e. the system is overrun).

    These file purges keep the /work filesystem available and performing at a more reliable and higher level. Some initial purge statistics for HCC resources are as follows:

    Crane: Total files: 36,306,639 Total amount: 274TB Oldest file removed was last accessed 1033 days ago.
    Sandhills: Total files: 14,192,087 Total amount: 62TB Oldest file removed was last accessed 1449 days ago.
    Tusker: Total files: 18,221,530 Total amount: 81TB Oldest file removed last last accessed at least 420 days ago.

    Long term storage is available for $60/TB/year on Attic (http://hcc.unl.edu/attic).


    HCC now purges (i.e. removes) files from /work that are over 6 months old. You may run

    hcc-purge

    to see which files (if any) are old enough to be removed at the next purging at any time. Purges will be performed each week on an ongoing basis moving forward. Scans for files to purge will be done twice a week. Please see

    https://hcc-docs.unl.edu/display/HCCDOC/Handling+Data

    for further details.

Pages