From amitoj at jlab.org Wed Jan 14 11:05:59 2026 From: amitoj at jlab.org (Amitoj Singh) Date: Wed, 14 Jan 2026 16:05:59 +0000 Subject: [lqcd-users] NOTICE: Migration of your data to a new Lustre file-system Message-ID: Dear LQCD JLab compute and storage users, As we get ready to deploy a new 4PB Lustre-filesystem based storage hardware we need to address the topic of migrating existing data with small files from /volatile and /cache storage areas to the new filesystem. Small files are a performance killer for the Lustre filesystem and will slow down the data migration process and so we would like to take the opportunity to bring this to your attention. Small files are defined as files that have sizes less than 1MB. As of now we plan to not migrate any small files to the new storage system. To mitigate this, we encourage everyone to please do the following: 1. Check your group/project small file count in /volatile and /cache using this following web pages -> link to volatile, link to cache 2. Do the following to reduce your small file count: * delete files that are no longer needed -OR- * compress small files into a single file -OR - * move small files to /work. If you have any questions or concerns, please do not hesitate to email me directly. Amitoj Singh, HPC Architect High Performance Data Facility Computational Sciences and Technology Division Thomas Jefferson National Accelerator Facility -------------- next part -------------- An HTML attachment was scrubbed... URL: From wmoore at jlab.org Fri Jan 16 13:53:49 2026 From: wmoore at jlab.org (Wesley Moore) Date: Fri, 16 Jan 2026 18:53:49 +0000 Subject: [lqcd-users] =?windows-1252?q?Scheduled_Power_Outage_=96_January?= =?windows-1252?q?_20th_=28Cluster_Shutdown_at_9=3A00_AM=29?= Message-ID: On January 20th, Facilities will be recommissioning the new PDUs in our data center. To ensure the safety of our clusters and all connected equipment, we will be powering off all systems at 9:00 AM. Impacted Services Include: * ifarm and qcdi interactive systems * JupyterHub * Batch compute clusters * Storage systems dependent on cluster availability * Supporting services and infrastructure Outage Details: * Start: January 20th at 9:00 AM * Work: PDU recommissioning * Estimated Duration: Up to 8 hours * Service Restoration: We will send an update once full production has resumed Thank you for your continued understanding and cooperation. On behalf of Scientific Computing Operations, Wesley -------------- next part -------------- An HTML attachment was scrubbed... URL: From wmoore at jlab.org Tue Jan 20 18:33:37 2026 From: wmoore at jlab.org (Wesley Moore) Date: Tue, 20 Jan 2026 23:33:37 +0000 Subject: [lqcd-users] =?windows-1252?q?Scheduled_Power_Outage_=96_January?= =?windows-1252?q?_20th_=28Cluster_Shutdown_at_9=3A00_AM=29?= In-Reply-To: References: Message-ID: All affected services are now expected to be back in full production. As a reminder, there will be up-coming ifarm reboots as part of continued maintenance. Thank you for your patience and cooperation. On behalf of Scientific Computing Operations, Wesley ________________________________ From: Wesley Moore Sent: Friday, January 16, 2026 1:53 PM To: Wesley Moore via Jlab-scicomp-briefs ; lqcd-users at jlab.org Subject: Scheduled Power Outage ? January 20th (Cluster Shutdown at 9:00 AM) On January 20th, Facilities will be recommissioning the new PDUs in our data center. To ensure the safety of our clusters and all connected equipment, we will be powering off all systems at 9:00 AM. Impacted Services Include: * ifarm and qcdi interactive systems * JupyterHub * Batch compute clusters * Storage systems dependent on cluster availability * Supporting services and infrastructure Outage Details: * Start: January 20th at 9:00 AM * Work: PDU recommissioning * Estimated Duration: Up to 8 hours * Service Restoration: We will send an update once full production has resumed Thank you for your continued understanding and cooperation. On behalf of Scientific Computing Operations, Wesley -------------- next part -------------- An HTML attachment was scrubbed... URL: