Frontera User News

Queues down on all Large Scale Systems Monday 15 February 2021

Posted by Matthew Edeker on Feb 15, 2021 11:00:35 AM

Due to the recent state of emergency in the City of Austin due to extreme weather and the impact it is having on the power in Austin, TACC is closing all queues on all Large Scale Systems. This includes Frontera, Stampde2, Lonestar5, Longhorn, and all other TACC resources. Queues will re-open as...

Updated on Feb 19, 2021 3:32:48 PM

All TACC large scale production systems have resumed full operations after powering down and load shedding to help with the Texas power grid emergency.

Original Posting

Due to the recent state of emergency in the City of Austin due to extreme weather and the impact it is having on the power in Austin, TACC is closing all queues on all Large Scale Systems. This includes Frontera, Stampde2, Lonestar5, Longhorn, and all other TACC resources. Queues will re-open as soon as possible when the emergency subsides.

Please submit any questions you may have via the TACC User Portal. 

Reminder: Frontera Texascale Days 2-8 February 2021

Posted by Tim Cockerill on Jan 29, 2021 1:03:50 PM

Frontera will be reserved for Texascale Days - very large scale runs of at least half-system to full-system size from Monday, February 2 through Monday, February 8. This event allows our researchers the opportunity to work on research problems that may not be possible at the smaller scale.  We have...

Updated on Jan 29, 2021 1:22:08 PM

Correction: Tuesday Feb 2 is the start date. My original post incorrectly had Monday Feb 2.

Original Posting

Frontera will be reserved for Texascale Days - very large scale runs of at least half-system to full-system size from Monday, February 2 through Monday, February 8. This event allows our researchers the opportunity to work on research problems that may not be possible at the smaller scale. 
We have not yet scheduled our next Texascale Days event. Watch for a forthcoming announcement if you have a research problem that will benefit from running at this scale.

Frontera Maintenance Scheduled on Tuesday 9 February 2021.

Posted by Alex Ferrier on Jan 28, 2021 8:53:26 PM

Frontera will not be available from 8 a.m. to 7:30 p.m. (CT) on Tuesday, 9 February 2021. System maintenance will be performed during this time.

Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting

Frontera will not be available from 8 a.m. to 7:30 p.m. (CT) on Tuesday, 9 February 2021. System maintenance will be performed during this time.

Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting

Stockyard Status Wednesday 27 January 2021

Posted by Matthew Edeker on Jan 27, 2021 11:03:51 AM

We're noticing performance issues with Stockyard and have closed queues on Frontera, Lonestar5, Stampede2, and Maverick2 while attempting to recover. Running jobs not using Stockyard (/work) will not be impacted.  Please submit any questions you may have via the TACC User Portal....

Updated on Jan 27, 2021 11:55:27 AM

Stockyard recovery process has been completed and queues have been opened. 

Original Posting

We're noticing performance issues with Stockyard and have closed queues on Frontera, Lonestar5, Stampede2, and Maverick2 while attempting to recover. Running jobs not using Stockyard (/work) will not be impacted. 


Please submit any questions you may have via the TACC User Portal.

Matlab unavailable on all TACC clusters (Frontera/Stampede2/Lonestar5/Maverick2) on Thurs 28 January 2021 due to license maintenance.

Posted by Alex Ferrier on Jan 26, 2021 6:02:52 PM

Matlab may be unavailable on all TACC clusters (Frontera/Stampede2/Lonestar5/Maverick2) on  on Thursday, January 28 from 5:00 PM - 6:00 PM CST.  The license service may be intermittently unavailable during this maintenance as licenses are updated for the 2021 calendar year. Please submit any...

Matlab may be unavailable on all TACC clusters (Frontera/Stampede2/Lonestar5/Maverick2) on  on Thursday, January 28 from 5:00 PM - 6:00 PM CST.  The license service may be intermittently unavailable during this maintenance as licenses are updated for the 2021 calendar year.


Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting

Emergency Network Maintenance Friday, 22 January 2021

Posted by Mark Brueschke on Jan 22, 2021 1:43:58 PM

Emergency firewall maintenance is being carried out, the process should be brief with all services restored quickly.


Please submit any questions you may have via the TACC User Portal.
https://portal.tacc.utexas.edu/tacc-consulting

Updated on Jan 22, 2021 2:53:54 PM

Firewall maintenance is complete as of 2:50 PM.

Original Posting

Emergency firewall maintenance is being carried out, the process should be brief with all services restored quickly.


Please submit any questions you may have via the TACC User Portal.
https://portal.tacc.utexas.edu/tacc-consulting

Upcoming TACC Training Short Courses - February and March 2021

Posted by Jason Allison on Jan 21, 2021 9:21:44 AM

We are pleased to announce the following training courses being held via webcast during the months of February and March 2021:  - Introduction to Linux - February 10th, 2021, 1pm - 4pm CT  - C++ for C Programmers - February 18th, 2021, 9am - 3pm CT  - MPL Object-Oriented Interface to MPI - February...

We are pleased to announce the following training courses being held via webcast during the months of February and March 2021:

 - Introduction to Linux - February 10th, 2021, 1pm - 4pm CT
 - C++ for C Programmers - February 18th, 2021, 9am - 3pm CT
 - MPL Object-Oriented Interface to MPI - February 25th, 2021, 9am - 12pm CT
 - Introduction to Machine Learning at TACC - March 5th, 2021, 9am - 3pm CT
 - Introduction to Deep Learning at TACC - March 12th, 2021, 9am - 3pm CT

Registration for these events closes at 12pm CT the day prior. To register and for more information please visit:

Please email jasona@tacc.utexas.edu if you have any questions.

TACC $WORK File System (Stockyard) is Offline - Thurs 14 Jan 2021

Posted by Alex Ferrier on Jan 14, 2021 9:13:50 PM

Stockyard ($work) is has experienced a failure and is unavailable right now. Admins are working to resolve the issue.


Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting


Updated on Jan 14, 2021 9:57:45 PM

The issue with the Stockyard filesystem has been resolved and queues have been opened.


Please submit any questions you may have via the TACC User Portal.

Original Posting

Stockyard ($work) is has experienced a failure and is unavailable right now. Admins are working to resolve the issue.


Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting


Frontera Texascale Days - 2-8 February 2021

Posted by Jesse Snead on Jan 14, 2021 5:06:02 PM

Frontera will be reserved for Texascale Days - very large scale runs of at least half system size (3,700 - 4,000 nodes) or full system size (7600 - 7900 nodes) Monday, February 2 through Monday, February 8. This event allows our researchers the opportunity to work on problems that may not be...

Frontera will be reserved for Texascale Days - very large scale runs of at least half system size (3,700 - 4,000 nodes) or full system size (7600 - 7900 nodes) Monday, February 2 through Monday, February 8. This event allows our researchers the opportunity to work on problems that may not be possible at the smaller scale. To be eligible to participate, your application must have already successfully run on at least 2,048 nodes in the “large” Queue on Frontera.  For access to the “large” queue, please submit a ticket with scaling data.

On the last day, Feb. 8, 2 hour blocks will be available for benchmarking efforts rather than production runs.  The other days, February 2 - February 7, will be set aside for production runs.  On these days, each project will have dedicated access to the number of nodes requested for a 24 hour period starting at 0900 CST.

To participate, please fill out and submit the form (https://forms.gle/oeTiBL9hPW6UyGZm9) before COB January 22.

Please submit any questions you may have through the TACC Consulting System or feedback form. 

https://portal.tacc.utexas.edu/tacc-consulting  

https://portal.tacc.utexas.edu/feedback

Frontera Maintenance Tuesday 12 2021

Posted by Mark Brueschke on Jan 5, 2021 1:06:02 PM

Frontera will not be available from 8:00 AM to 10:00 PM (CST) on Tuesday, 12 January 2021. System maintenance will be performed during this time.

Updated on Jan 12, 2021 10:28:17 PM

Frontera is back in full production. 

Original Posting

Frontera will not be available from 8:00 AM to 10:00 PM (CST) on Tuesday, 12 January 2021. System maintenance will be performed during this time.

TACC Outage December 31 2020

Posted by Alex Ferrier on Dec 31, 2020 9:58:40 PM

TACC's datacenter has had a power event that's impacting multiple core systems. Queues on all of the Large Scale Systems will remain down while assessing the impact and recovering. Admins are working to restore services as soon as possible. Please submit any questions you may have via the TACC User...

Updated on Jan 1, 2021 6:24:06 PM

The Lonestar5 issue related to the power event has been resolved and queues have been opened. Please contact us via a ticket if you see any further problems on the system. Thank you, TACC LSS team

Updated on Jan 1, 2021 12:14:25 PM

Frontera, Stampede2 and other services are back on line as of 12 noon (CST) on Friday, January 1 2021. Lonestar5 remains offline at this time as we work to restore it.

Original Posting

TACC's datacenter has had a power event that's impacting multiple core systems. Queues on all of the Large Scale Systems will remain down while assessing the impact and recovering. Admins are working to restore services as soon as possible.

Please submit any questions you may have via the TACC User Portal.
https://portal.tacc.utexas.edu/tacc-consulting

-TACC Team

Network Performance Issue

Posted by Garland Whiteside on Dec 22, 2020 1:55:41 AM

TACC staff are investigating a network performance issue which is affecting login process to TACC systems. Staff are working to resolve the problem.

Updated on Dec 22, 2020 2:22:41 AM

TACC Staff has concluded the work on the network performance issue.  The issue has been resolved. 


Thanks, 
TACC Staff.

Original Posting

TACC staff are investigating a network performance issue which is affecting login process to TACC systems. Staff are working to resolve the problem.

Frontera Maintenance Tuesday 8 December 2020

Posted by Mark Brueschke on Dec 1, 2020 12:39:09 PM

Frontera will not be available from 8:00 AM on Tuesday, 8 December 2020 to 8:00 AM (CST) on Wednesday, 9 December 2020. System maintenance will be performed during this time.

Please submit any questions you may have via the TACC User Portal.
https://portal.tacc.utexas.edu/tacc-consulting

Updated on Dec 9, 2020 8:03:04 AM

Frontera is back in production as of 8:00 AM Wednesday, 9 December 2020.

Original Posting

Frontera will not be available from 8:00 AM on Tuesday, 8 December 2020 to 8:00 AM (CST) on Wednesday, 9 December 2020. System maintenance will be performed during this time.

Please submit any questions you may have via the TACC User Portal.
https://portal.tacc.utexas.edu/tacc-consulting

Frontera Texascale Days - Full System Runs 9-14 December 2020

Posted by Tim Cockerill on Nov 18, 2020 9:18:40 AM

Frontera will be reserved for Texascale Days - very large scale runs of at least 3,800 nodes up to full system - to be held immediately following the December system maintenance, Wednesday December 9 through Monday December 14. This event allows our researchers the opportunity to work on problems...

Updated on Dec 8, 2020 8:44:07 AM

UPDATE - Frontera Texascale Days Extended for COVID-19 Simulations


Frontera Texascale Days are now scheduled through 9am Wednesday, December 16, 2021. Frontera will return to normal service at that time. 
We anticipate the next Texascale Days event to be held in late January or early February. Watch for the announcement that will have details on when your research team can apply to participate.

Best regards,
Tim Cockerill
Director of User Services

Original Posting

Frontera will be reserved for Texascale Days - very large scale runs of at least 3,800 nodes up to full system - to be held immediately following the December system maintenance, Wednesday December 9 through Monday December 14. This event allows our researchers the opportunity to work on problems that may not be possible at the smaller scale. To be eligible to participate, your application must have already successfully run on at least 2,048 nodes in the Large Queue on Frontera.If you would like to participate in Texascale Days, please submit a ticket and include a brief description of the experiment you would like to run and how many nodes would be required.  If you have not yet run on at least 2,048 nodes and would like to request access to the large queue, include this in your ticket. This should include your own strong or weak scaling results from Frontera on up to 512 nodes.If you have already run in the large queue on at least 2,048 nodes, you are qualified to request a Texascale reservation for running at 3,800 nodes or more.

Code Performance and Scaling for Frontera Proposals - December 9th, 2020 1:00 PM - 2:30 PM CT

Posted by Jason Allison on Nov 27, 2020 4:43:14 PM

We are pleased to announce the Code Performance and Scaling for Frontera Proposals webinar being held on December 9th, 2020 from 1:00 PM to 2:30 PM CT. During this event we will describe the opportunities for requesting access to both Frontera and Longhorn, provide guidance on how to determine...

We are pleased to announce the Code Performance and Scaling for Frontera Proposals webinar being held on December 9th, 2020 from 1:00 PM to 2:30 PM CT.

During this event we will describe the opportunities for requesting access to both Frontera and Longhorn, provide guidance on how to determine which opportunity is most appropriate for your project, and describe effective scaling techniques found among successful proposal requests.

Registration for this event closes at 12 PM CT on December 9th, 2020. To register please visit:

Please email jasona@tacc.utexas.edu if you have any questions.

Frontera Pathways and LSCP Allocations now open for proposals through November 20

Posted by Jesse Snead on Nov 13, 2020 10:34:49 AM

We are pleased to announce the fourth Frontera Pathways and Frontera Large Scale Community Partnership (LSCP) allocation tracks will remain open through November 20. Resources available for request include Frontera and its GPU subsystem, as well as the Longhorn GPU resource.  Pathways These are...

We are pleased to announce the fourth Frontera Pathways and Frontera Large Scale Community Partnership (LSCP) allocation tracks will remain open through November 20. Resources available for request include Frontera and its GPU subsystem, as well as the Longhorn GPU resource


Pathways

These are small allocations (between 10,000-250,000 SUs per year) to science teams with a strong scientific justification for access to a leadership-class computing resource, but who have not yet demonstrated code readiness to effectively utilize such a resource.


LSCP

These are allocations (between 25,000-1,000,000 SUs per year, up to 3 years) where we can’t strictly characterize the set of experiments to run within a 12-month allocation period, but that line up with a large team or community, a large instrument, or other large NSF investment.


Detailed descriptions of all four Frontera allocation tracks can be found on the Frontera Allocations web page.


How to Submit an Allocation Request

Please begin by reading the Allocations Policy & Submission Guidelines that provides information regarding who is eligible to apply, minimum and maximum allocation sizes, and an outline of the information that is required to be included in your proposal.


After reading the allocations policy and guidelines, you may determine that a startup allocation would be beneficial to do code performance and benchmarking to be incorporated into your Pathways or LSCP proposal. Instructions for obtaining a Startup allocation are included on the policy & guidelines web page.


Please submit any questions you may have through the TACC Consulting System or feedback form.


https://portal.tacc.utexas.edu/tacc-consulting 

https://portal.tacc.utexas.edu/feedback


Thanks,

Tim Cockerill

Director of User Services

Texas Advanced Computing Center

The University of Texas at Austin


/work Filesystem Status Wednesday 11 November 2020

Posted by Matthew Edeker on Nov 11, 2020 10:19:42 AM

The /work filesystem is currently unavailable due to a server problem, administrators are working to resolve the issue.


Please submit any questions you may have via the TACC User Portal.

Updated on Nov 11, 2020 11:10:21 AM

The issue has been resolved and queues are being opened. 

Original Posting

The /work filesystem is currently unavailable due to a server problem, administrators are working to resolve the issue.


Please submit any questions you may have via the TACC User Portal.

Frontera Status Thursday 5 November 2020

Posted by Matthew Edeker on Nov 5, 2020 11:55:25 AM

Frontera's /scratch1 filesystem is unavailable right now and we are working with the vendor to get it restored.


Please submit any questions you may have via the TACC User Portal.

Updated on Nov 5, 2020 3:46:55 PM

Frontera is back in full production. 

Original Posting

Frontera's /scratch1 filesystem is unavailable right now and we are working with the vendor to get it restored.


Please submit any questions you may have via the TACC User Portal.

Frontera Emergency Maintenance Monday October 26

Posted by Mark Brueschke on Oct 26, 2020 2:23:36 PM

Frontera's /scratch3 filesystem is partially unavailable. TACC staff is working with the vendor to resolve this issue.

Updated on Oct 28, 2020 4:59:07 PM

Frontera's /scratch3 filesystem is back in full production.

Updated on Oct 27, 2020 3:57:46 PM

The /scratch3 filesystem on Frontera remains offline while the controller verifies the storage arrays, which based on current progress should complete some time tomorrow afternoon. We hope to have the filesystem back available to users by 5:00PM tomorrow afternoon, Wednesday Oct 28th.

Updated on Oct 26, 2020 4:31:14 PM

The failed controller on /scratch3 continues to be worked on and DDN is investigating, but it might be offline until tomorrow. For now, we have cancelled all running jobs using /scratch3 and held pending jobs running out of /scratch3 to prevent them from failing. The queues have been re-opened for those users running out of /scratch1 or /scratch2. Users of /scratch3 should avoid submitting new jobs until the filesystem access has been restored.

Original Posting

Frontera's /scratch3 filesystem is partially unavailable. TACC staff is working with the vendor to resolve this issue.

Containers @ TACC - 6 November, 2020 9am-3pm

Posted by Jason Allison on Oct 21, 2020 4:00:39 PM

We are pleased to announce the Containers @ TACC Training being held on November 6th, 2020 from 9am to 3pm CT. Software containers are an important common currency for portable and reproducible computing. Learn best practices on building, using, and sharing Docker and Singularity containers in this...

We are pleased to announce the Containers @ TACC Training being held on November 6th, 2020 from 9am to 3pm CT.

Software containers are an important common currency for portable and reproducible computing. Learn best practices on building, using, and sharing Docker and Singularity containers in this hands-on workshop. Also learn how to run those containers on TACC HPC systems, including MPI and GPU aware containers.

Registration for this event closes at 5pm CT on November 4th, 2020. To register please visit:

Please email jasona@tacc.utexas.edu if you have any questions.

Frontera Maintenance 20 October 2020

Posted by David Littrell on Oct 6, 2020 4:31:16 PM

Frontera will not be available from 8 a.m. to 7:30 p.m. CST on Tuesday, 20 October 2020. System maintenance will be performed during this time.

Updated on Oct 20, 2020 7:30:36 PM

Frontera maintenance has been extended until 10:30pm CST.

Original Posting

Frontera will not be available from 8 a.m. to 7:30 p.m. CST on Tuesday, 20 October 2020. System maintenance will be performed during this time.

/work Filesystem Status Thursday 10 September 2020

Posted by Matthew Edeker on Sep 10, 2020 2:21:55 PM

The /work filesystem is currently unavailable due to a failed switch, TACC staff are working to replace the switch and restore access to the filesystem.


Please submit any questions you may have via the TACC User Portal

https://portal.tacc.utexas.edu/tacc-consulting

Updated on Sep 10, 2020 2:52:11 PM

The /work filesystem switch has been repaired and /work is available now on all systems.


Please submit any questions you may have via the TACC User Portal.

https://portal.tacc.utexas.edu/tacc-consulting

Original Posting

The /work filesystem is currently unavailable due to a failed switch, TACC staff are working to replace the switch and restore access to the filesystem.


Please submit any questions you may have via the TACC User Portal

https://portal.tacc.utexas.edu/tacc-consulting