cscs-ci run eiger-spack-psmp

@RMeli It's weird that you could not trigger the cscs-ci run. I checked the list of globally trusted CI users and your username is included.

The CI test eiger-spack-psmp fails, because the branch dla-pxpotri does not include the most recent changes in the master.

It's weird that you could not trigger the cscs-ci run.

Indeed. I tried to change the capitalization of my username, but doesn't seem to be the problem. I'll have a look.

@RMeli There is a time out when pulling the CP2K image as sqsh file. That error did not appear before and is most likely not caused by this PR. Maybe, it is related to the migration eiger to eiger.alps?

It seems that there were JFrog issues around the time I triggered the build. I will try again, otherwise inquire internally (there were a few issues this morning due to the migration, but they should have been fixed).

cscs-ci run eiger-spack-psmp

It seems that the container download needs now 9 min and thus the time limit of 10 min for the test job is not sufficient.

Looks like the problem is with the entry point, and possibly with the switch from Sarus (on eiger) to the Container Engine (on eiger.alps).

I'm a bit time constrained this week, I'm writing you an e-mail to organize a meeting.

cscs-ci run daint-spack-psmp

cscs-ci run daint-toolchain-psmp

cscs-ci run eiger-spack-psmp

cscs-ci run eiger-spack-psmp

cscs-ci run daint-spack-psmp

Is this PR ready to be merged?

@oschuett I did fix the previous CI issues so for me it was ready to merge. But as discussed by e-mail, the changes introduced in #4297 make all DLA-Future tests fail due to an issue seemingly related to MPICH CH4:OFI.

While it's unfortunate that we can't run DLA-Future in the CI right now, I'm sure you tested this PR manually.
So, I'm leaning towards including it in the release nevertheless.

Should be fine for the release, since PR #4316 excludes DLAF for the time being.