Upcoming service outages, November 19 through Nov 22.
Hi, everyone.
I mentioned this in the last note, but want to make sure we’re clear on what’s going to be up and running vs. what’s not next weekend. I’ll try to keep it succinct.
As Building 240 occupants are aware, the building will lose power on Saturday, November 20. This includes the data center. There’s a bunch of work we’re scheduling to coincide with this and take advantage of the forced outage. Please note, this notice serves to provide outage info specifically for general CELS services in both the General Computing Environment (GCE) and the Legacy MCS network. Announcements related to LCRC, ALCF, and other systems in the room will be handled by their respective support teams, and those timelines may differ from what I mention below.
Friday, November 19
8:30AM: Most GCE resources will go offline. GCE login nodes will remain operational as a pass-through to Legacy or other systems. During this outage window, we will be powering down, removing, and moving the primary file server for GCE from building 240 to building 386, aka the Enterprise Data Center (EDC). Legacy and all other systems will remain online for this period.
3:00PM (or sooner if possible): GCE will come back online with the file server in its new location, no longer encumbered by 240 power work. Most GCE resources will remain up through the remainder of this outage window.
5:00PM: MCS Legacy environment, and GCE compute nodes housed in 240 will go offline. This includes the legacy login nodes, compute001.mcs.anl.gov, compute002.mcs.anl.gov, and compute004.mcs.anl.gov. In GCE, this affects compute-03 through compute-06.cels.anl.gov. compute-01 (also runs nomachine) and 02 will remain up through this work, as will the GCE Home nodes (homes-01 and 02).
Out of an abundance of caution, we will be taking the CELS Account Management system (accounts.cels.anl.gov) offline during this work starting at 5PM. We are working towards being able to keep this up during outages like this.
This will not affect your ability to connect to GCE resources, but you will not be able to make changes to your account during this window.
All desktops and other computers in the building will lose power later in the evening as the building’s power is shut down.
Monday, November 22
8:30AM: Legacy and GCE services taken down for the power work will be brought back up. Most services will return quickly. We will send an all clear when this is complete, we expect before 10AM.
None of this work will affect CELS email services, web services, or any hosted services such as Slack.
I’ll send a final reminder of this work on Thursday of next week. Thanks, and have a great weekend!