Daily Bulletin Archive

April 25, 2013

For the past two days, Yellowstone utilization has been hovering around 50%. The system is ready and waiting for user jobs.

If you are encountering any issues with getting started on Yellowstone or having Yellowstone complete your jobs, please contact cislhelp@ucar.edu.

April 25, 2013

No Scheduled Downtime: Yellowstone, Geyser_Caldera, HPSS, GLADE, Lynx

April 23, 2013

NCAR's Mesa Lab data center will be undergoing significant electrical work during the semi-annual power-down scheduled for Saturday, April 20, from 6 a.m to 6 p.m. While NWSC will not be directly affected, users will notice several effects of this maintenance period:

* During this outage, HPSS will be taken out of service to eliminate the possibility of errors or confusion while the Mesa Lab tape libraries are powered down.

* Some YubiKeys and CryptoCARDs will not work, including many of those belonging to UCAR staff, while the NCAR RADIUS server at Mesa Lab is out of service. The RADIUS server is slated to be returned to service as early as possible after the maintenance work has been completed.

* CISL's GridFTP service (and GlobusOnline) will be unavailable from 5 p.m. April 19 through 12 p.m. April 21. The GridFTP service is scheduled to be migrated to NWSC, but not until after the power-down.

* CISL's license server, which enables software including Matlab, IDL, and Fluent to be used on Yellowstone, Geyser, and Caldera, is also in the process of being migrated to NWSC. Licensed software may also not be usable from 5 p.m. April 19 through 12 p.m. April 21.

* Finally, CISL user support relies on a number of services that will also be down, including the ExtraView ticket system, Notifier, email, and the CISL web site. The CISL Help Desk phone line, 303-497-2400, will still be available for urgent situations, but note that Saturday is outside of normal business hours.

Aside from these Mesa Lab-housed services, Yellowstone, Geyser, Caldera, and GLADE will continue to operate normally during the power-down. Users are encouraged to submit jobs to be run over the April 20 weekend, though users may need to work around the HPSS outage.

April 22, 2013

The Yellowstone compute nodes were returned to service around 12:30 a.m., April 19, after an unplanned outage. More than 200 queued jobs began running on Yellowstone minutes after the queues were opened.

The outage was due to severe weather in the Cheyenne area that caused power bumps at the NWSC facility. NWSC did not experience a total outage, but the power bumps, which began about 9 a.m. MT, April 18, were enough to trigger protective measures in the NWSC data center.

During the day, the UPS and battery backup systems performed as designed, which allowed other components of the Yellowstone environment to ride through the power bumps. These systems include the Yellowstone login nodes, Geyser, Caldera, GLADE, and HPSS. Those systems remained available to users throughout the day.

Any users with jobs that were running when the compute nodes went down can contact cislhelp@ucar.edu about getting refunds for those jobs. Please include the job number or numbers.

April 12, 2013

While Yellowstone has regularly been averaging 90% utilization on weekdays, it has room to handle more jobs. In particular, overnight and on weekends, Yellowstone has been emptying its queues and showing lower utilization in the early morning hours and on Sundays and early Mondays.

Users are encouraged to factor this usage pattern and submit jobs late in the day or late on Friday. Such jobs are likely to be run overnight or over the weekend. Note that these times might be good opportunities to take advantage of the economy queue.

April 11, 2013

The Mirage cluster and the GLADE resource at the Mesa Lab were taken down and retired on April 1, as announced previously. Files that remained in the Mesa Lab GLADE scratch and project spaces were removed and are no longer available.

Files that were in /glade/home were backed up and are being retained off-line for six (6) months. Users who need to access those files should contact cislhelp@ucar.edu.

During the transition period, the new scratch space was identified as /glade/nwsc_scratch; it is now just /glade/scratch. If you still have “/glade/nwsc_scratch” in a script, you will need to change it to “/glade/scratch” to use the new space.

For more information regarding the new GLADE file spaces for Yellowstone users, see http://www2.cisl.ucar.edu/resources/glade.

April 5, 2013

Consulting walk-in services will be closed this week as all of our user consultants will be busy in SEA conference. If you want consultation over the telephone we will create a consulting ticket for you and a consultant will return your calls at the earliest opportunity. You may continue to submit tickets as usual to cislhelp@ucar.edu.

April 5, 2013

NCAR/CISL invites NSF-supported university researchers in the atmospheric, oceanic, and related sciences to submit large allocation requests for the Yellowstone system by 4 April 2013. All requesters are strongly encouraged to review the instructions before preparing their submissions. No late requests will be accepted.

These requests will be reviewed by the CISL High-performance computing Advisory Panel (CHAP), and there must be a direct linkage between the NSF award and the computational research being proposed. Please visit http://www2.cisl.ucar.edu/docs/allocations for more university allocation instructions and opportunities.

Allocations will be made on Yellowstone, NCAR's 1.5-petaflops IBM iDataPlex system; the data analysis and visualization clusters (Geyser and Caldera); the 11-petabyte GLADE disk resource, and the High Performance Storage System (HPSS) archive. Please see https://www2.cisl.ucar.edu/resources/yellowstonefor more system details.

For the Yellowstone resource, the threshold for Small University Allocations has been increased to 200,000 core-hours; researchers with such needs can apply for small allocations at any time. A Small Allocation is also recommended for researchers new to Yellowstone, in order to conduct benchmarking and test runs prior to submitting a large allocation request.

April 2, 2013

All users are reminded that the Mirage cluster and the GLADE resource at the Mesa Lab will be taken down and retired on April 1. This means users must have completed all data transfer by Sunday, March 31.

After March 31, all data will be deleted from the old Mesa Lab GLADE file spaces, including /glade/scratch, /glade/proj2, and /glade/proj3. All files in those areas will be removed on April 1 and will no longer be available.

Files in /glade/home have been backed up and will be retained off-line for 6 months. Users should contact cislhelp@ucar.edu to access those files.

March 28, 2013

No Scheduled Downtime: Yellowstone, Geyser_Caldera, HPSS, NWSC GLADE, DAV, ML GLADE, Lynx