Open bedroge opened 5 months ago
We're now seeing the same issue for our EESSI-extend module (see #578), which is a bundle.
bedroge@x86-64-amd-zen3-node2 ~ $ ls -la /cvmfs/software.eessi.io/versions/2023.06/software/linux/x86_64/amd/zen3/software/EESSI-extend/2023.06-easybuild/
total 0
drwxrwxr-x 2 bedroge bedroge 42 Jun 7 11:54 .
drwxr-xr-x 3 bedroge bedroge 31 Jun 7 11:54 ..
# so it looks like the software was successfully removed, but you the old easybuild subdir is still there somehow:
bedroge@x86-64-amd-zen3-node2 ~ $ ls -la /cvmfs/software.eessi.io/versions/2023.06/software/linux/x86_64/amd/zen3/software/EESSI-extend/2023.06-easybuild/easybuild/
total 62
dr-xr-xr-x 3 bedroge bedroge 4096 May 7 13:41 .
drwxrwxr-x 2 bedroge bedroge 42 Jun 7 11:54 ..
-r--r--r-- 1 bedroge bedroge 2754 May 7 13:41 EESSI-extend-2023.06-easybuild-easybuild-devel
-r--r--r-- 1 bedroge bedroge 7976 May 7 13:28 EESSI-extend-2023.06-easybuild.eb
-r--r--r-- 1 bedroge bedroge 22875 May 7 13:41 easybuild-EESSI-extend-2023.06-20240507.134135.log.bz2
-rw-rw-r-- 1 bedroge bedroge 19051 May 7 13:41 easybuild-EESSI-extend-2023.06-20240507.134135_test_report.md
dr-xr-xr-x 4 bedroge bedroge 4096 May 7 13:41 reprod
I still don't have a clue why it only happens for some easyconfigs...
When trying to rebuild Python, hatchling, and Python-bundle-PyPI in #546, we ran into weird permission issues for both hatchling and Python-bundle-PyPI. It's not clear yet what's causing it, but it seems to happen for (Python) bundles that include extensions that not only install files to
lib
, but also tobin
.. The removal step seems to work fine and successfully removes the existing installation, but in the build phase the extension will suddenly see the oldbin
directory again (with read-only permissions), and fail with errors like:I've tried a lot of possible workarounds in both #546 and #555:
chmod -R u+w
) instead of removing the existing installation dirhatchling
) instead of only removing the installation dir of the particular versionEASYBUILD_READ_ONLY_INSTALLDIR
before starting the buildNone of them solved the issue, though. So, in the end, I opted for working around it by adding write permissions to the affected installation directories on the Stratum 0, and then the rebuilds completed successfully.