Daily Bulletin Archive

Aug. 8, 2017

The Cheyenne cluster and GLADE project space — /glade/p — will be unavailable from 8 a.m. until 10 a.m. MDT on Tuesday, August 8,  to allow CISL staff to perform important system maintenance. Also, users will be logged off and locked out of the Cheyenne login nodes during the maintenance window. Yellowstone, Geyser, and Caldera users may experience interruptions or errors accessing /glade/p files  during this maintenance.

A system reservation will be put in place on batch jobs Monday afternoon on Cheyenne. Jobs submitted with specified job times that overlap the reservation period will remain on hold until the system is restored to service. Running jobs that have not finished when the outage begins may be killed and need to be resubmitted after the system is restored to service. Users are advised to take this into account before submitting long jobs after 8 p.m. on Monday.  

We apologize for any inconvenience this might cause. Users will be informed via the CISL Notifier service when the system is returned to production.

Jul. 30, 2017

The Cheyenne cluster will be unavailable from 8 a.m. MDT on Saturday, July 29, until 5 p.m. Sunday, July 30, to allow CISL staff to perform important system tests.

A system reservation will be put in place on batch jobs Friday afternoon. Jobs submitted with specified job times that overlap the reservation period will remain on hold until the system is restored to service. Running jobs that have not finished when the outage begins may be killed and need to be resubmitted after the system is restored to service.

We apologize for any inconvenience this might cause. Users will be informed via the CISL Notifier service when the system is returned to production.

Jul. 27, 2017

Cheyenne users can now run a “qpeek” script to check on the progress of their running jobs. The script inspects a specified job’s stdout log by default. Documentation added to Managing and monitoring jobs describes how to check stderr instead, and how to merge stdout and stderr in a job script so that qpeek can check both while the job is running.

Jul. 20, 2017

A number of users have contacted the CISL Help Desk to ask about how to change their default login shell. The quickest way to do that—for both Cheyenne and Yellowstone—is through our Systems Accounting Manager (SAM). In the “Login Shell” section of your preferences screen, select a system and then click the “Edit” button to make your change. It may take several hours for the change to take effect.

What else can you do with SAM? See this page to find out.

Jul. 18, 2017

CISL resumed enforcing the 45-day /glade/scratch purge policy on June 1. The policy was suspended during the recent transition to the new, larger GLADE scratch system. Users’ files that are not accessed (for example: modified, read, or copied) within the 45 days following June 1 will be deleted automatically on July 15.

Also on June 1, the temporary,  read-only /glade/scratch_old file space was removed and will be repurposed. Please see GLADE File Spaces for more information about using the GLADE centralized file service.

Jul. 18, 2017

Registration is now open for this summer’s free ACI-REF Virtual Residency Workshop on “How to Be a More Effective Research Computing Facilitator.” Participants can attend remotely via videoconference or in person at the University of Oklahoma from Sunday, July 30, to Saturday, August 5.

The workshop is intended for those who want to be more effective at helping researchers and educators use research cyberinfrastructure. Participants are expected to give a five-minute presentation on a project that they intend to carry out at their home institution after the workshop ends. They also are asked to participate in a second summer workshop and in Virtual Residency conference calls during the next two academic years. See the application form for more information.

Jul. 17, 2017

NCAR supercomputing, data storage, and archive systems at the NCAR-Wyoming Supercomputing Center (NWSC) will be powered down during a scheduled power outage July 14-18 for mechanical and electrical system testing and certification at the facility in Cheyenne.

The Cheyenne, Yellowstone, Geyser, Caldera, Erebus, GLADE, and HPSS systems will be unavailable from 6 a.m. Friday, July 14, until 8 a.m. Tuesday, July 18. Associated test systems such as Laramie also will be unavailable.

A system reservation will be put in place on batch jobs the afternoon of Thursday, July 13. After the system reservation is in place, jobs submitted with specified job times that overlap the beginning of the reservation period will remain on hold until the system is restored to service. Running jobs that have not finished when the systems are powered down will be killed and will need to be resubmitted when the systems are restored to service.

Users will be informed via the CISL Notifier service when the systems return to production and advised to check on the status of jobs that were on hold.

Jul. 7, 2017

Where and how to start jobs on the Cheyenne and Yellowstone HPC systems and the Geyser and Caldera clusters have been common questions as more and more users begin working on Cheyenne.

The short answer is to log in to Cheyenne to run jobs on Cheyenne, and log in to Yellowstone to submit jobs to run on Yellowstone, Geyser, or Caldera.  The separate logins are necessary because the systems have different job schedulers and operating systems.

We are working to provide a more streamlined solution for Cheyenne users to run data analysis tasks on Geyser and Caldera. Users who need to transfer existing, automated workflows to Cheyenne should contact cislhelp@ucar.edu to discuss options.

See these CISL web pages for detailed documentation:

Jul. 4, 2017

The CISL Help Desk and Consulting office will be open with reduced coverage on Monday, July 3, and then closed for the July 4 holiday. Users who need assistance are encouraged to email cislhelp@ucar.edu.

Calls to 303-497-2400 will be answered on Monday but some users may need to leave voicemail messages. On Tuesday, calls will be forwarded to staff at the NCAR-Wyoming Supercomputing Center.

Jul. 3, 2017

Cheyenne’s economy queue is being enabled for a limited time to encourage greater use of the system. Jobs using the economy queue will be charged a discounted rate of 30% compared to the regular queue.  

Jobs submitted to the regular and economy queues will have the same priority due to the current PBS Pro configuration, so all users are encouraged to use the economy queue while it is available. Its availability will be re-evaluated as Cheyenne’s usage increases.

For more information on Cheyenne’s queues see Job-submission queues and charges.

Pages