Will you actually want anything on tigergpu? We don't use GPUs.

Beyond that, unless the hardware is very exotic, the only thing that matters is probably the operating system. I think they all have access to /tigress, right? Can you try simply running the Perseus-built stack on new-Tiger? I think that'll probably Just Work.

Ah, you are correct:

[lam3@tigercpu ~]$ module load rh/devtoolset/6

[lam3@tigercpu ~]$ source /tigress/HSC/LSST/stack_perseus/loadLSST.bash

Note: use devtoolset-6 in conjunction with this stack.

[lam3@tigercpu ~]$ setup lsst_distrib

[lam3@tigercpu ~]$ eups list -s lsst_distrib

 16.0+11 current w_2018_28 setup

And a trial processCcd.py ran fine.

As for tigergpu...I have no idea if we'll ever want to use it, but the above worked there too, so the point is moot . So I suppose this can be marked invalid (or altered to just make some updates to the documentation here)?

I like the idea of this being a doc update.

Hey Lauren, do you mind taking a look at https://github.com/lsst-dm/dm_dev_guide/pull/249?

I'm deliberately shortening the instructions here to just the essentials, in part because I'm no longer au fait with what's actually being used by the group in Princeton. If there's more that you think should be added, perhaps you could suggest some text?

Hey Lauren MacArthur — gentle reminder!

So sorry this slipped off my radar!  The text all looks good to me, but I just tried setting up the stacks and they are giving me:

lam3@tiger2-sumire:~ $ eups list -s lsst_distrib

 16.0+36 current w_2018_33 setup

lam3@perseus:~ $ eups list -s lsst_distrib

 16.0+36 current w_2018_33 setup

whereas lsst-dev has:

lauren@lsst-dev01:~ $ eups list -s lsst_distrib

 16.0+41 current w_2018_34 setup 

Is the auto-update not working properly?

The build on new-Tiger is failing with

$ cat /tigress/HSC/LSST/stack3_perseus_20171107/stack/miniconda3-4.3.21-10a4fa6/EupsBuildDir/Linux64/daf_persistence-16.0-3-g3806c63+6/daf_persistence-16.0-3-g3806c63+6/tests/.tests/PropertySet_2.failed 

tests/PropertySet_2

Running 1 test case...

*** No errors detected

which seems less than entirely helpful, but it doesn't seem to be a problem with these docs.

I propose to merge the text on this ticket and file a new one to investigate the build failure. Ok with you?

A-ok with me.

Thanks Lauren; merged. Eli Rykoff will be glad to hear that the instructions are now not obviously incorrect!

Follow-up ticket for build failures in DM-15593.