Retiring aging compute nodes

As noted when we released the new compute nodes (see https://mcssys.wordpress.com/2018/12/14/mcs-environment-compute-nodes/), we’re going to be turning off the compute nodes these machines supplanted. Our goal was to turn these off before the break, but the best laid plans of mice and men, yada yada yada. Still, what must be done must be done, so, at COB on Thursday, Feb 21, we’ll turn off the following computers.
crush.mcs.anl.gov
crank.mcs.anl.gov
steamroller.mcs.anl.gov
grind.mcs.anl.gov
churn.mcs.anl.gov
trounce.mcs.anl.gov
thrash.mcs.anl.gov
vanquish.mcs.anl.gov
octagon.mcs.anl.gov
If you have any crontabs on these hosts, you will want to migrate them to one of the new compute nodes.
Also, if you have any data in /sandbox that is not duplicated elsewhere, shame on you. /sandbox is ephemeral, not backed up, and can go away at any time. Still, we’re a nice bunch so we’re letting you know you’ve got a couple of days to get any unique data you’re storing in /sandbox to a place that’s more reliable. If you need help with any of this, let us know at [email protected].
As a side note, the host octopus.mcs.anl.gov will stay online through the end of this calendar year, as it’s the only 32-bit machine in the environment. If we have a need for such a host after December, we’ll work with those who have the need to come up with the proper solution.
Thanks!