-
Notifications
You must be signed in to change notification settings - Fork 14
Reload EESSI-extend when switching to a different EasyBuild version #79
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
New job on instance
|
So, this is indeed completely broken. Let's try to fix it now... |
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
New job on instance
|
load_easybuild_module.sh
Outdated
|
||
# EESSI-extend checks for the EB version being used. | ||
# If EESSI-extend is already loaded, we need to reload it in order to reevaluate the checks. | ||
module is-loaded EESSI-extend && module update |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This doesn't control the order in which modules are reloaded, so may not fix anything.
I would unload
and then load
EESSI-extend?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It should, as we have already switched the EB version at this point, and then it will simply reload both that EB version and EESSI-extend (order doesn't matter). But maybe it's cleaner anyway to only reload EESSI-extend, so I'll change it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just doing a module load EESSI-extend
again (if it's loaded) should work:
$ source /cvmfs/software.eessi.io/versions/2023.06/init/bash
$ ml EESSI-extend
{EESSI 2023.06} $ echo $EASYBUILD_CUDA_SANITY_CHECK_ERROR_ON_FAILED_CHECKS
1
{EESSI 2023.06} $ ml EasyBuild/4.9.4
The following have been reloaded with a version change:
1) EasyBuild/5.1.1 => EasyBuild/4.9.4
{EESSI 2023.06} $ echo $EASYBUILD_CUDA_SANITY_CHECK_ERROR_ON_FAILED_CHECKS
1
{EESSI 2023.06} $ ml EESSI-extend
{EESSI 2023.06} $ echo $EASYBUILD_CUDA_SANITY_CHECK_ERROR_ON_FAILED_CHECKS
Previous build failed because bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
New job on instance
|
Now I get a really weird syntax error:
|
Let's try the bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
New job on instance
|
That sounds like you're hitting a Maybe related to: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm
When loading EESSI-extend, it will automatically load the latest EB version. The EESSI-extend module does some EB version checks, and based on the version it may set certain EB env vars. If you then switch to an older EB version, these checks are not reevaluated, which may lead to a broken EB setup and errors like:
This does not contain a fix yet, I first want to reproduce the issue with a simple example.