Flux, the Flux Hadoop cluster, and their storage systems (/home, /scratch, and HDFS) will be unavailable starting at 8 p.m. Saturday, January 2, 2016, returning to service on Wednesday, January 6.
REMINDER: Any submitted HPC jobs must be completed by 8 p.m., Saturday, Jan. 2. The maximum run-time for any job on Flux is 28 days.
During the outage, the following projects have been planned:
- OS and supporting software updates for the cluster. This will be a minor update to the currently installed RedHat version (RHEL 6.7)
- Cluster management software will be updated
- Updates to some default software versions and retirement of some older software packages and/or versions
- Architectural improvements to the InfiniBand network
- Deployment of new InfiniBand-to-Ethernet gateways
- Software updates to the Hadoop cluster
Status updates will be posted on the ARC-TS Twitter feed.
Any questions can be directed to email@example.com.