Radon
-
The Radon cluster has reached the end of its useful lifespan and will be decommissioned on Monday, April 2nd, 2018. Radon has served the Purdue research community for a great many years and has undergone radical changes during that time. However, du...
-
All Research Computing systems suffered an unplanned outage Saturday, March 24th, 2018 at 8:15pm EDT due to a widespread power failure in the area. Thanks to diligent efforts all night and today by many teams across ITaP, all computational clusters h...
-
New Windows Network Drive (SMB) Access
All Windows network drives (SMB/CIFS access) for scratch filesystems on all clusters and home directories have moved! You should change your mapped network drives to connect to: \\scratch.my_cluster_name_here.rcac.purdue.edu\my_cluster_name_here Or f...
-
Unscheduled Depot Outage on Compute Clusters
The servers providing access to Data Depot from Brown, Conte, Halstead, HalsteadGPU, Radon, Rice, Scholar, and Snyder suffered a partial failure. Many nodes in these clusters temporarily lost access to Depot. Jobs accessing files on Depot may have pa...
-
Purdue University will be observing a holiday break from December 23 - January 2. During this time, Research Computing services will continue to be available, but all staff will be on leave. Critical system outages will be dealt with should they occ...
-
Radon, Hathi, and WSC Hadoop Cluster Maintenance
The Hathi, Radon, and WSC Hadoop clusters will be unavailable beginning at Friday, November 3rd, 2017 at 5:00pm EDT, for scheduled electrical power maintenance. The clusters will return to full production by %enddatetime%. During this time, some of t...
-
Unscheduled Outage in Math Data Center
At approximately 2:00pm EDT on Tuesday, September 5th, 2017, the Math building data center lost some power feeds which supply the Conte, Halstead, HalsteadGPU, Hathi, and Radon clusters. Scheduling on these has been paused until we can be sure power...
-
After a long review, Research Computing has determined it is necessary to alter the scratch storage purge policy on all systems. Effective August 28, 2017, all scratch storage systems will begin purging files which have not been accessed (for either...
-
Hathi, Radon, and Specialized Cluster Maintenance
Update message: After performing necessary repairs, Radon has been returned to service. -- Previous message: After consulting with vendor support, we have determined that Radon has experienced a failure in its network hardware. Parts and and a vendo...
-
Removal of netcdf and hdf5 system-level library installations
Research Computing has begun removing several libraries installed at the system-level that should be provided by the module command. These libraries include netcdf, hdf5, and several related packages. This change should have limited impact as module...
-
Engineers have restored failed core servers back to a functional state. Data Depot is up and running as normal and job scheduling resumed. Should you encounter any lingering issues please let us know at rcac-help@purdue.edu Original Message Some core...
-
Clusters to complete transition to hierarchy modules
The transition to hierarchy modules has completed today, May 9th. If any of your job scripts still have old module names the module load will no longer work so be sure to double check your PBS job scripts and output for any old module names. We have...
-
Emergency Security Patching of RCAC Clusters
Due to a recent security vulnerability, the Carter, Halstead, Hammer, Radon, Rice, Scholar, and Snyder clusters will have their operating system upgraded to a newer version during February 2, 2017 5:00pm - March 2, 2017 5:00pm EST. Unlike other cl...
-
Radon to switch to hierarchical modules
Radon has been converted to hierarchical modules. Old style module load commands should still continue to function until the end of the Spring semester, however you will be given a warning message and a suggestion on how to correct your scripts. Once...
-
Job scheduling paused on Radon
Job scheduling was paused on Radon between 6 pm and 7 pm this evening. Node monitoring processes marked most nodes offline around 6 pm, preventing new jobs from starting. System engineers cleared the fault in the node monitoring, and nodes came back...
-
Security vulnerability patch impacts debugging
Due to a recently found vulnerability in the Linux Kernel (known as the Dirty-COW vulnerability), an emergency patch has been applied on the cluster nodes. This patch is necessary to avoid exploitation of the vulnerability. Unfortunately, the patch i...
-
Home Filesystem Maintenance - All Clusters
Conte has been returned to normal operations as well now. This concludes the home directory maintenance on all systems. Update: September 27, 2016 11:55pm All systems other than Conte have been successfully returned to normal operations with the ne...
-
Software stack changes and upgrades
During the Home Filesystem Maintenance - All Clusters maintenance on September 27th, several upgrades and changes will be made to the software stack on the clusters. Changes will include updates to the default version of the Intel compiler and associ...
-
Degraded performance of several systems
We have seen a significant wave of these events this morning, September 21. For the most part, this wave seems to have been linked to a storage problem that has been resolved. However, we are implementing new monitoring and response procedures toda...
-
Update As of 5:50 pm, Tuesday, 16 Aug 2016, the Radon cluster has been returned to service and is fully operational. Thank you for your patience. Update Due to unanticipated conflicts between the upgraded scheduler and our network configuration, the...