Add possibility to ignore certain hooks on local modules#530
Add possibility to ignore certain hooks on local modules#530trz42 merged 4 commits intoEESSI:2023.06-software.eessi.iofrom casparvl:ignore_eessi_hooks_on_local_modules
Conversation
…'t want to be prevented from loading local CUDA modules because of the EESSI hook. See #523
|
Instance
|
…tensions in /home/casparl/eessi/...
|
Just to confirm, I've tested again after Alan's latest changes. On a module from EESSI: On a local CUDA module: So still works as intended :) I think this is ready to be merged, but... since it's my own PR it's not really up to me :D @ocaisa did you want @trz42 do a final check? They already merged it in NorESSI. @trz42 it should be easy to test: just initialize the EESSI environment, load the CUDA module from EESSI (you should get the warning). Then, load any CUDA module from a local module environment, and that should not give you the warning. |
|
This has been already deployed to NESSI. A test (not yet taking locations as definable in #371 into account) succeeded (see below). Therefore, we can trigger the builds and deploy updated |
|
bot: build repo:eessi.io-2023.06-software arch:x86_64/generic |
Updates by the bot instance
|
Updates by the bot instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
The |
|
Hm, no wait, it can't, since I didn't sync this branch with develop (yet). I'm actually not sure why it's redeploying this file :\ I do know that @boegel should have deployed it in his PR... |
|
And yet if I look in the bot's tarball, I see: That's the updated version of the |
The bot merges your PR with current version of target branch, so you're essentially also doing the "build" for the changes in #451 here, indeed... |
|
Ah ok. Well, that's fine. Then we can just deploy :) As long as we understand what's happening :D |
|
bot: build repo:eessi.io-2023.06-software arch:x86_64/generic |
Updates by the bot instance
|
Updates by the bot instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
We don't want to be prevented from loading local CUDA modules because of the EESSI hook. See #523
This PR makes sense that the CUDA hooks are only applied if the module being loaded comes from the EESSI prefix.
I.e. before this change:
After this change:
Fixes #523