Infrastructure News Type: Outage Full

Affected Infrastructure: anvil.purdue.access-ci.org, anvil-gpu.purdue.access-ci.org

Start Date: December 11, 2025, 12:00 p.m.

End Date: December 11, 2025, 10:51 p.m.

UPDATE: December 11, 2025 05:51 PM ET As of 5:51pm ET, the maintenance work on Anvil has been completed and job scheduling has been resumed. If you encounter any issues post maintenance, please contact ACCESS Help Desk (https://support.access-ci.org/help-ticket). Original Post: The Anvil system will be unavailable on December 11th, 2025, from 7:00 AM to 6:00 PM ET for scheduled maintenance. During this maintenance, we will perform NVIDIA driver upgrades and rebuild Slurm with PMIx. Any Slurm jobs requesting a walltime that would extend past Thursday, December 11th at 7:00 AM ET will not start and will remain in the queue until after maintenance is completed. If you have any questions, please submit a ticket through the ACCESS Help Desk (https://support.access-ci.org/help-ticket).

Posted: March 20, 2026 • Author: Nannan Shan

Infrastructure News Type: Reconfiguration

Affected Infrastructure: jetstream2.indiana.access-ci.org

Start Date: December 2, 2025, 12:00 p.m.

End Date: December 3, 2025, 2:00 a.m.

On December 2, Jetstream2 engineers will be performing updates to Jetstream2's networking infrastructure to enable DNS resolution over IPv6. There is no expected downtime or impact to users for this maintenance. If you have any questions, please contact Jetstream2 Support by emailing help@jetstream-cloud.org

Posted: March 20, 2026 • Author: Jennifer Armstrong

Infrastructure News Type: Degraded

Affected Infrastructure: aces.tamu.access-ci.org

Start Date: November 18, 2025, 7:40 p.m.

End Date: November 18, 2025, 9:55 p.m.

TAMU ACES had a storage outage at 1:40 pm ~ 3:55 pm CST, Nov 18, 2025. ACES storage is now back to normal.

Posted: March 20, 2026 • Author: Tsung-i Huang

Infrastructure News Type: Outage Full

Affected Infrastructure: launch.tamu.access-ci.org

Start Date: October 25, 2025, 2:00 p.m.

End Date: October 27, 2025, 4:00 a.m.

Updated: 10:25 pm CDT, Oct 26, 2025. Launch cluster maintenance has completed. All queued jobs are running. The TAMU Launch cluster will be unavailable from 9a CDT Saturday October 25 to 11pm CDT Sun October 26 for regular and storage maintenance.

Posted: March 20, 2026 • Author: Tsung-i Huang

Infrastructure News Type: Outage Full

Affected Infrastructure: xdmod.access-ci.org

Start Date: October 16, 2025, 2:00 p.m.

End Date: October 17, 2025, 2:00 p.m.

The ACCESS XDMoD portal will temporarily be unavailable on 10/16/2025, from approximately 09:00 EST until the following day, 10/17/2025, 09:00 EST. The service will be completely unavailable for routine infrastructure updates. An update will be posted once the update is complete. We apologize for any inconvenience this might cause. UPDATE: The update has completed successfully and the service is available again.

Posted: March 20, 2026 • Author: Conner Saeli

Infrastructure News Type: Outage Partial

Affected Infrastructure: jetstream2-gpu.indiana.access-ci.org

Start Date: October 7, 2025, 12:00 p.m.

End Date: October 8, 2025, 12:00 a.m.

Jetstream2 GPU resources will undergo routine monthly maintenance on October 7, 2025 from approximately 11AM to 11PM UTC (convert to local time (https://dateful.com/convert/utc?t=11am)). This maintenance will include firmware updates and other necessary changes to improve GPU performance. During this maintenance window users may experience intermittent delays and unavailability of their GPU resources. As always, we will do our best to keep disruptions to a minimum during this time. We strongly advise preserving your work prior to October 7 by: - Safely shutting down any active processes or jobs - Backing up essential data outside of Jetstream2 Please note: Jetstream2 CPU, Large Memory, and Storage resources will not be affected. If you have any questions or concerns, please contact Jetstream2 Support at help@jetstream-cloud.org.

Posted: March 20, 2026 • Author: Jennifer Armstrong

Infrastructure News Type: Outage Full

Affected Infrastructure: launch.tamu.access-ci.org

Start Date: September 27, 2025, 11:00 p.m.

End Date: September 30, 2025, 5:00 p.m.

The Launch cluster maintenance has been extended to 12pm CDT Tuesday September 30 The TAMU Launch cluster will be unavailable from 6p CDT Saturday September 27 to 10a CDT Monday September 29 for regular maintenance and the 9/28 data center partial power outage. Quotas will also be enabled for user/project scratch directories during this maintenance window

Posted: March 20, 2026 • Author: Francis Dang

Infrastructure News Type: Outage Partial

Affected Infrastructure: aces.tamu.access-ci.org

Start Date: September 27, 2025, 11:00 p.m.

End Date: September 29, 2025, 3:00 p.m.

The TAMU ACES cluster will be operating at reduced capacities with a subset of nodes powered down during data center power maintenance from 6p CDT Saturday September 27 to 10a CDT Monday September 29.

Posted: March 20, 2026 • Author: Francis Dang

Infrastructure News Type: Degraded

Affected Infrastructure: stampede3.tacc.access-ci.org, ranch.tacc.access-ci.org

Start Date: September 27, 2025, 5:06 a.m.

End Date: October 1, 2025, 5:00 p.m.

Final Update: 10/1/2025 14:15 CDT All ACCESS available TACC resources are now back in full production. If you have questions about other TACC resources you can refer to User Updates here: https://tacc.utexas.edu/news/user-updates/ ---- Update: 09/30/2025 17:00 CDT Most TACC services have been restored since Saturday’s power outage. Thanks to the dedicated work of the team, all large systems are in production, with the exception of Frontera which is running a limited number of jobs while performance issues on one of the scratch filesystems are addressed. You may see some services at less than 100% throughput while replacement parts trickle in for components that were damaged. If anything appears offline that should not be, please submit a ticket here: https://tacc.utexas.edu/portal/tickets Thank you for your continued patience while the admins continue to work to restore services. ---- UPDATE: 9/29/25 Stampede3 has partially recovered from an emergency maintenance. /corral is not available and queues will be operating with reduced node counts as we work to bring the system back up to 100%. ---- Original - 9/27/25 The Texas Advanced Computing Center facilities experienced a site wide power outage at 6:12 am CDT. At this time, admins are on site and working to restore all systems to production.

Posted: March 20, 2026 • Author: Virginia Trueheart

Infrastructure News Type: Reconfiguration

Affected Infrastructure: anvil.purdue.access-ci.org, anvil-gpu.purdue.access-ci.org

Start Date: September 23, 2025, 1:00 p.m.

End Date: September 23, 2025, 9:00 p.m.

Update: As of 10:30 am EDT, September 23, 2025, we have finished Anvil Open OnDemand Dashboard upgrade. We are thrilled to announce the launch of the new version of our re-designed dashboard for Open OnDemand on Anvil! This upgrade is designed with a more intuitive interface and a host of more powerful new features to enhance your experience on Anvil. Please check the original post for the overview of all the amazing new features AND Get Started Today Access the new dashboard by visiting https://ondemand.anvil.rcac.purdue.edu. We’ve prepared a comprehensive guide (see “New Features” page in the new dashboard) to help you navigate and make the most out of the new features. Have questions? Need assistance? Require more features? Our support team is here to help! Contact us through ACCESS Help Desk at https://support.access-ci.org/help-ticket (https://support.access-ci.org/help-ticket**). We’d love to hear your feedback as you explore the new gateway. Your input is invaluable in shaping the future of our HPC services. Thank you for being part of this journey. Let’s push the boundaries of discovery together! Original post: The Open Ondemand service for Anvil will be unavailable from Tuesday, September 23 at 9:00am EDT, 2025 to Tuesday, September 23 at 5:00pm EDT, 2025. During the maintenance, Anvil team will perform a reconfiguration to the Open Ondemand dashboard for Anvil which will upgrade the current dashboard to a new version with new features listed below. What’s New on the dashboard v2? - New UI design: Brand new UI design to present a more modern look. - Anvil AI Partition Status: Adding partition status check for the new Anvil AI partition. - Cluster Status and Node Status: Allow users to have a glance of overall Anvil cluster status or dive deep into the status for specific compute node. - New Announcement Widget: Now you can view past announcements with a scroll! - New My Jobs page: Adding more features to My Jobs page. - New Job page: View or control your specific jobs more easily through the job page. - New Performance Metrics page: Now you can view your job performance metrics on Anvil within any time span. What will impact you? - All Slurm jobs on Anvil (including jobs that have already submitted through Open Ondemand before this maintenance) will continue and NOT be impacted. - All functions related to Open Ondemand including login will be unavailable during the maintenance. Anvil Open Ondemand service will return to full production by Tuesday, September 23 at 5:00pm EDT, 2025. Please submit a ticket through ACCESS Help Desk at https://support.access-ci.org/help-ticket (https://support.access-ci.org/help-ticket**) if you have any questions or suggestions.

Posted: March 20, 2026 • Author: Guangzhen Jin

Infrastructure News Type: Outage Full

Affected Infrastructure: expanse.sdsc.access-ci.org, expanse-gpu.sdsc.access-ci.org

Start Date: September 22, 2025, 1:00 p.m.

End Date: September 22, 2025, 9:00 p.m.

Dear Expanse User, The Expanse Slurm maintenance has been completed and we have released the reservation. The system is running submitted jobs per available resources and priority. Thanks SDSC User Services >>>> Dear Expanse User, The maintenance on Expanse is taking longer than expected and we are extending the maintenance reservation to 6PM(PT). We will update once the maintenance is complete. Thanks SDSC User Services Staff >>>>> Dear Expanse User, We will have a maintenance on Expanse 6AM-Noon (PT), September 22, 2025. During this maintenance we will reconfigure the Slurm scheduler to remove some nodes that were moved to a different cluster. We have a reservation in place to prevent jobs from running during this maintenance. The "squeue" output will show "ReqNodeNotAvail, Reserved for maintenance" for jobs that do not fit in the time period before the maintenance begins. These jobs will run after we release the maintenance reservation. Thanks SDSC User Services Staff

Posted: March 20, 2026 • Author: Mahidhar Tatineni

Infrastructure News Type: Outage Partial

Affected Infrastructure: aces.tamu.access-ci.org, launch.tamu.access-ci.org

Start Date: September 19, 2025, 11:00 p.m.

End Date: September 21, 2025, 5:00 a.m.

The TAMU ACES and Launch clusters will be be operating at reduced capacities with a subset of nodes powered down during data center power maintenance from 6p Friday September 19 to midnight Sunday September 21.

Posted: March 20, 2026 • Author: Francis Dang

Infrastructure News Type: Outage Partial

Affected Infrastructure: anvil.purdue.access-ci.org, anvil-gpu.purdue.access-ci.org

Start Date: August 18, 2025, 4:00 p.m.

End Date: August 20, 2025, 10:00 p.m.

Anvil is experiencing a problem with new user and allocation propagation. Our engineers are working on the fix, and will keep this updated. The problem has been fixed on 5 pm.

Posted: March 20, 2026 • Author: Xiao Liu

Infrastructure News Type: Outage Partial

Affected Infrastructure: aces.tamu.access-ci.org, launch.tamu.access-ci.org

Start Date: August 2, 2025, 1:00 a.m.

End Date: August 2, 2025, 11:00 a.m.

The TAMU campus network will be undergoing maintenance from 8p CDT Aug.1 to 6a CDT Aug. 2. The TAMU ACES, FASTER, and Launch clusters will be inaccessible to ACCESS users for at least the first 20 minutes of the maintenance window. During the remainder of the maintenance, there may be intermittent connectivity issues for accessing the TAMU clusters. The network maintenance will not impact running jobs on the TAMU clusters.

Posted: March 20, 2026 • Author: Francis Dang

Infrastructure News Type: Outage Partial

Affected Infrastructure: anvil-gpu.purdue.access-ci.org

Start Date: July 31, 2025, 9:00 p.m.

End Date: August 1, 2025, 5:00 p.m.

Update: The Anvil AI nodes (Nvidia H100 GPUs) have been resumed at 11:10am EST. Thank you for your patience. Original: The Anvil AI nodes (Nvidia H100 GPUs) are currently powered off due to ongoing cooling issues in the data center. Facilities has confirmed that the cooling system will not be restored until sometime tomorrow, and the H100 GPUs will remain offline until it is safe to bring it back online. Job scheduling remains paused, but file access is still available during this downtime. We now anticipate service restoration sometime tomorrow (Friday, August 1). We will provide additional updates as more information becomes available or by 12:00pm EST. Thank you for your patience and understanding.

Posted: March 20, 2026 • Author: Nannan Shan

Infrastructure News Type: Outage Full

Affected Infrastructure: xdmod.access-ci.org

Start Date: July 30, 2025, 6:30 p.m.

End Date: July 31, 2025, 5:00 p.m.

The ACCESS XDMoD portal will temporarily be unavailable today, 07/30, from approximately 13:30 EDT until tomorrow, 07/31, 12:00 EDT. The service will be completely unavailable for routine infrastructure updates.

Posted: March 20, 2026 • Author: Conner Saeli

Infrastructure News Type: Outage Partial

Affected Infrastructure: expanse.sdsc.access-ci.org, expanse-gpu.sdsc.access-ci.org, expanse-ps.sdsc.access-ci.org

Start Date: July 18, 2025, 1:00 p.m.

End Date: July 19, 2025, 1:00 p.m.

Dear Expanse User, We are currently seeing some issues with the Lustre metadata server and that will cause filesystem write issues. Please pause/hold any Lustre based jobs and we will update once the issue is resolved. Thanks SDSC User Services

Posted: March 20, 2026 • Author: Mahidhar Tatineni

Infrastructure News Type: Outage Partial

Affected Infrastructure: xdmod.access-ci.org

Start Date: July 17, 2025, 2:00 p.m.

End Date: July 18, 2025, 2:00 p.m.

ACCESS XDMoD will be upgraded to version 11.0.2 on Thursday, July 17 at approximately 10:00 EDT. Various data in ACCESS XDMoD may be unavailable during the upgrade. Service is expected to be fully restored within 24 hours. Once the upgrade is started, release notes will be available at https://xdmod.access-ci.org/#main_tab_panel:about_xdmod?Release%20Notes

Posted: March 20, 2026 • Author: Aaron Weeden

Infrastructure News Type: Outage Partial

Affected Infrastructure: aces.tamu.access-ci.org

Start Date: July 11, 2025, 2:55 p.m.

End Date: July 11, 2025, 6:00 p.m.

UPDATE: The degraded storage server was recovered around 11:50a. We are continuing to monitor the ACES filesystem for any further issues. We are currently seeing degradation on one of the Lustre storage servers. This is leading to slow filesystem access and impacting the responsiveness of the Slurm job scheduler. We will update once the issue is resolved.

Posted: March 20, 2026 • Author: Francis Dang

Infrastructure News Type: Outage Partial

Affected Infrastructure: expanse.sdsc.access-ci.org, expanse-gpu.sdsc.access-ci.org, expanse-ps.sdsc.access-ci.org

Start Date: July 8, 2025, 1:00 a.m.

End Date: July 8, 2025, 5:00 p.m.

>>> Update Dear Expanse User, We resolved the Lustre OSS issues yesterday and have been monitoring the filesystem for any further issues. Separately this morning (7/9/2025) we had a short disruption of Slurm job submissions due to a down system service and that has also been resolved. Thanks SDSC User Services >>> Dear Expanse User, We are currently seeing high load and timeouts on one of the Expanse Lustre filesystem object storage servers (OSSs). This might lead to access issues on files that are striped onto storage targets on this OSS. We are looking into the issue and will update once it is resolved. Thanks SDSC User Services Staff

Posted: March 20, 2026 • Author: Mahidhar Tatineni