240 Data Center Outage May 8-11, GCE Linux Upgrades April 24-26

Hi, everyone! I want to let you know about a couple of outages/upgrades that are happening.

First the easy one — we’ll be rebuilding any GCE Ubuntu 18.04 machines into Ubuntu 22.04 starting next week. On April 24 we’ll be doing homes-01.cels.anl.gov, compute-01.cels.anl.gov, and compute-02.cels.anl.gov. On April 26 we’ll do homes-02.cels.anl.gov. In all of these cases, the machines will be wiped clean. If you have any data in /scratch on any of these machines, you’ll want to put it somewhere else. (Always treat /scratch like it’s temporary storage, because it is. It is not backed up.)

Now the big one — the building 240 data center will experience a complete shutdown of cooling and a partial shutdown of power beginning the morning of May 8, through the end of day May 11. This is to support the cooling upgrade project as part of bringing Aurora to life.

All computers in the data center will be powered down. The systems administrators of the various systems in the room will send specific notices to each user community detailing how they’ll be handling the outage. In general, you should expect things to shut down on Sunday evening (May 7), and return to service at some time on Friday May 12.

Most systems in GCE will be unaffected by this as they are hosted in a different data center, however some compute nodes that live in building 240 will be powered off due to this work.

All Legacy storage will be offline. As such, homes-legacy.mcs.anl.gov will not be accessible. GCE login/home nodes will continue to operate normally.

I will send a reminder closer to the shutdown.

Thanks!