|
|
News
Dedicated Time for /nobackupp8 Completed
13 Dec, 2018
The dedicated time for the /nobackupp8 filesystem has completed. Previous notification The Lustre filesystem /nobackupp8 will be placed in dedicated time for software and hardware patching from 11:00 a.m. to 5:00 p.m. (Pacific Time) on...
New PBS Feature: Reserving Skylake Nodes for "At-Scale" (Wide) Jobs
06 Dec, 2018
A new PBS queue, called sky_wide, is now available to help you run very wide (at-scale) jobs on Electra's Skylake nodes by ensuring that at least once a week, the jobs in the queue will have enough available nodes to run. Each week, jobs...
NASA Brings its Science and Supercomputing Advances to SC18
05 Nov, 2018
Some of NASA's most exciting achievements enabled by the agency's supercomputers are being showcased at the international Supercomputing Conference (SC18), scheduled Nov. 12–16, 2018 in Dallas, Texas. This year marks the 30th anniversary of the...
Shift File Transfer Failures
01 Nov, 2018
At approximately 12:43 a.m. Pacific Time on November 1, 2018, the drive holding metadata for Shift file transfers failed. At 9:55 a.m., new metadata was redirected to a different drive. If you have used Shift to transfer data recently, please note:...
Lou is Now Available
31 Oct, 2018
The issue described in the previous notification has been resolved, and Lou is online and available. We do not believe there is any data loss resulting from this problem. However, if your Lou home filesystem is one of the following, and you...
Dedicated Time for All HECC Systems Has Completed
30 Oct, 2018
Dedicated time for all systems, networks, and filesystems in the HECC enclave has been completed. All HECC systems are now available. As always, if you have questions, please contact the Control Room at (800) 331-8737, (650) 604-4444, or...
Upgrade to Shift Automated Transfer Components
22 Oct, 2018
During the dedicated time for all HECC systems (scheduled for October 26-29), the Shift transfer tool will be upgraded with several enhancements, which are listed below. The upgraded Shift tool will be available when the systems come back online...
Reminder: Electra Dedicated Time Scheduled for October 17-29
16 Oct, 2018
Electra will be placed into dedicated time beginning Wednesday, October 17 at 7:00 a.m., and continuing through the HECC enclave outage, which is scheduled to end Monday, October 29 at 11:59 p.m. (Pacific Time). The Electra dedicated time is...
Dedicated Time on PFEs, LFEs Scheduled for October 10 and 11
09 Oct, 2018
The Pleiades and Lou front-end systems (PFEs and LFEs) will be placed in dedicated time for software patching, according to the following schedule: Wed. Oct. 10, noon-2:00 p.m. (PT): pfe20, pfe21, pfe22, pfe23, lfe1, and lfe3 Thur. Oct. 11,...
Reminder: SLES 11 Software Packages in /nasa Directory Will Be Removed on Oct. 6
03 Oct, 2018
All HECC systems at the NAS facility have been upgraded to SUSE Linux Enterprise Server 12 (SLES 12), and the SLES 11 software is scheduled for removal on Saturday, October 6. This affects all software that is available through...
New Standard Billing Unit (SBU) Rates Effective October 1
30 Sep, 2018
Starting October 1, 2018, a new common Standard Billing Unit (SBU) is being used for allocating and tracking computing time usage at both the NASA Advanced Supercomputing (NAS) Facility and the NASA Center for Climate Simulation (NCCS). Impact to...
Update: Significant Problems on /nobackupp2 Resolved
27 Sep, 2018
The /nobackupp2 filesystem has returned to its expected operation and performance. A configuration change in our InfiniBand network resolved the issue. Unfortunately in the process of troubleshooting the issue, a number of PBS jobs were...
Check Files Written to Lou During Unscheduled Outage on Monday, September 24
25 Sep, 2018
An unscheduled outage of the Lou mass storage system occurred on Monday, September 24, from approximately 3:30 p.m. - 7:30 p.m. (Pacific Time). The outage may have caused some file transfer issues if you were using any of the /s2* filesystems....
Electra Expanded with Additional Skylake Nodes
15 Aug, 2018
Electra has been expanded with the addition of 576 Intel Xeon Gold 6148 "Skylake" nodes, bringing the total number of Skylake nodes to 1,728. This expansion brings Electra's theoretical peak performance to 6.55 petaflops—approaching that of...
Automounting of Filesystems (Possible Action Needed)
07 Aug, 2018
Filesystems are now automatically mounted and unmounted as requested either by batch jobs or on demand, and are automatically unmounted after a period of inactivity (usually 30 minutes). This has been in effect for several months, and has reduced...
Follow-up Information on July 12-13 Electra Outage
13 Jul, 2018
A problem with Electra that began on July 12 at approximately 11:35 p.m. Pacific Time was resolved this morning, July 13. The problem was caused by an electrical panel fire in another building at NASA Ames, which in turn caused Electra's circuit...
Lou Data Analysis Nodes and Front-Ends lfe1 and lfe2 Are Available
14 Jun, 2018
The Lou data analysis nodes (ldans) and front-ends lfe1 and lfe2 are now available. The software patch has been applied to fix the SLES 12 filesystem issues. All SLES 12 ldans are available through PBS. lfe1 and lfe2 running SLES 12 are also...
Issue with Lou SLES 12 Filesystem
25 May, 2018
Due to a SLES 12 filesystem issue, all Lou SLES 12 systems are unavailable. We are working with our vendor to resolve this issue. In the meantime, please note: You can use the following Lou front-end nodes (LFEs), which have been downgraded to...
Training Materials for "Compiler Directive-based GPU Programming in C and Fortran" Now Available
30 Apr, 2018
Training materials are now available for the webinar "Compiler Directive-based GPU Programming in C and Fortran," which was held on Wednesday, April 25, 2018. You can download a PDF of the slides presented, and follow along with an accompanying...
Charging SBUs on Sandy Bridge GPU Nodes
04 Apr, 2018
Starting April 5, 2018, projects utilizing Sandy Bridge GPU nodes are charged at an SBU rate of 3.64 (twice the current Sandy Bridge rate). To utilize GPU nodes in PBS, specify the queue k40 with model type san_gpu. For users who may not be...
|
