Please Note: Actual shutdown scheduled to begin Monday, 5/22 at 9pm (this event listing cannot span days).
UPDATE3: Scheduling and storage online. ATLAS not yet up. Clean-up of small issues. Contact firstname.lastname@example.org for any services marked operational that still present problems.
UPDATE2: FIXED Home directory storage issue causing slow/no login and stalled logins. Working on issue.
UPDATE1: Power-up has begun. An all-clear email will be sent once complete. Follow progress at: https://status.rc.fas.harvard.edu/
Summary: The MGHPCC (Holyoke) data center, one of our primary data centers, will be performing its annual full power shutdown and maintenance Tuesday, May 23rd, 2017. This will necessitate our powering down all resources at MGHPCC/Holyoke at 9PM on May 22nd. All services should be back online by noon on Wednesday 5/24.
This event will affect all jobs and most services. Running jobs will be stopped, pending jobs will remain pending. Storage and other resources hosted at MGHPCC will be unavailable.
This timeline is contingent upon MGHPCC completing their work according to their current schedule. We will keep the community posted on any changes as they occur.
POWER DOWN: Begins 9PM Monday 5/22/17
POWER OUT: All day Tuesday 5/23/17
POWER UP SCENARIO 2: RC will begin the power-up procedures Wednesday morning 5/24 with expected completion by noon (current ETA is 10AM).
The following devices/resources will not be available for the duration of the outage.
- Odyssey cluster
- regal scratch filesystem (/n/regal)
- holylfs Lustre filesystem
- pan2, pan3 storage systems
- CEPR Secure Environment
- aagfs01 storage system
- kuanglfs Lustre storage system (/n/kuanglfs)
- holyflash (SSD Lustre filesystem)
- datamover01 (high speed data transfer)
- any other device resident at MGHPCC