Messages & Announcements

  • 2012-09-14:  HCC Position Opening
    Category:  General Announcement

    The Holland Computing Center (HCC) is a large High Performance Computing facility supporting the NU system. HCC is looking to add a motivated, talented individual to work on ongoing international research collaborations such as the CMS (Compact Muon Solenoid) project. This is one of the particle accelerator experiments located at the Large Hadron Collider near Geneva, Switzerland. HCC seeks to hire a Systems Integrator with the following duties:

    Provide systems support for research computing, integrating distributed software components for building a robust infrastructure to support interdisciplinary research projects. Work with systems architect to create and combine components which integrate into a larger distributed framework for research support. Work in new development as well as refactoring and maintenance of legacy systems. Criminal background check will be conducted. Excellent benefits including staff/dependent scholarship program. Applicant review begins Sept 20. View requisition 120725 at https://employment.unl.edu for details and to apply. UNL is committed to a pluralistic campus community through affirmative action, equal opportunity, work-life balance, and dual careers.


  • 2012-08-17:  Sandhills open, some assembly required
    Category:  General Announcement

    Sandhills is now available. The filesystems were not changed, but the OS and the scheduler were. Please read the FAQ at http://hcc.unl.edu/sandhills/faq.php , and let us know if you have any difficulty.


    As mentioned in the previous announcement email, Prairiefire has been retired and 'merged' with Sandhills. The new Sandhills cluster is now online and available for use. There are some differences between the old and new environment, most notably the OS upgrade to Scientific Linux 6.3 and the change to SLURM as a job scheduler. -Please- read the FAQ at http://hcc.unl.edu/sandhills/faq.php for more details on the changes and how they may affect you.

    Unfortunately, while bringing the new Sandhills cluster online we discovered an issue with the storage used for /home. The issue is bad enough that normal file operations like copying were very slow and prone to failure.

    At this time we have not found a solution (or to be honest a cause) of this behavior and have decided to bring Sandhills online with the old /home fileserver as it was before. Thus, your home directories should look and feel just like they did before Prairiefire was retired. If you have logged in within the past 24 hours and find files in /home missing let us know at hcc-support@unl.edu and we will assist.

    We hope to resolve this filesystem issue as quickly as possible, at which point the new fileserver will be put in place. For what it's worth, this new fileserver is identical to the one serving Tusker's /home, which to date has worked well and given us few if any problems. The issues we're having on the Sandhills version are a complete surprise to us. For now we would rather have users able to log in and get work done than hold up everything while we sort out the problems.

    As this is a major system update there will undoubtedly be a few rough edges. We will do our best to resolve these quickly. Please use "module avail" to check for available software packages -- if you don't find an expected package installed, please let us know. While this amount of change will result in some level of disruption, it was required: the OS update was long overdue, and the scheduler was increasingly inadequate to meet evolving demands.

    Again - please read the FAQ at http://hcc.unl.edu/sandhills/faq.php, and if you have questions email hcc-support@unl.edu.

  • 2012-07-31:  PrairieFire retirement and Sandhills reconfiguration
    Category:  General Announcement

    PrairieFire and SandHills require a shutdown for substantial changes in two weeks (Aug. 14). Several significant improvements are planned: the OS will be upgraded, the Maui/torque scheduler will be replaced, and the 2 systems will be fully merged with the name "prairiefire" retired. Existing binaries and scripts may work immediately; some may require recompiling or editing. Updated FAQs will be available before these changes are enacted. If you have any concerns, please contact hcc-support@unl.edu.


    Several clusters sharing the name "PrairieFire" have served us well over the last 9 years, but, as announced at various user meetings last spring, the name will soon go into a well-earned retirement. August 14th the newer cluster SandHills, already sharing several components with PrairieFire, and the parts of PrairieFire that remain useful, will be merged into a single system. After August 14th prairiefire will no longer be maintained as a separate resource. This will coincide with several other important changes.

    The operating system is overdue for an upgrade; it will be updated to SL6.2 (equivalent to RHEL6). We will also be implementing the SLURM resource manager in place of TORQUE/Maui. Existing TORQUE submit scripts will run with minimal changes. FAQs will be updated on the HCC website soon. While existing executables will likely run, recompiling may be required. The current scheduling stack's failings have been mitigated by vigilance and a growing list of ad hoc solutions. SLURM (https://computing.llnl.gov/linux/slurm/overview.html) promises to improve this situation considerably, based upon our test implementation and the experience of LLNL and other prominent users. All filesystems will be unaffected by this change; they were already integrated previously.

    While these changes are necessary, even a successful upgrade can result in unexpected effects. Please contact us immediately if you encounter any trouble, and we'll work to make the transition as smooth as possible.

  • 2012-06-26:  Tusker downtime, July 2, 10am-noon
    Category:  Maintenance

    On Monday, July 2, from 10am to noon, Tusker will be unavailable in order to perform maintenance on the /home filesystem.

    No running jobs should be affected, but jobs may be delayed in the three days leading up to the maintenance window. Any jobs which cannot complete before the maintenance window will be held until work is complete.


    On Monday, July 2, from 10am to noon, Tusker will be unavailable in order to perform maintenance on the /home filesystem.

    No running jobs should be affected, but jobs may be delayed in the three days leading up to the maintenance window. Any jobs which cannot complete before the maintenance window will be held until work is complete.

  • 2012-06-15:  PRAIRIEFIRE: Unplanned outage
    Category:  System Failure

    The /util network filesystem crashed and had to be restarted this morning. If you had jobs running on PRAIRIEFIRE please log on and check to see if they need to be resubmitted.


    The /util network filesystem crashed and had to be restarted this morning. If you had jobs running on PRAIRIEFIRE please log on and check to see if they need to be resubmitted.

Pages