Closed hexylena closed 9 years ago
@erasche, @bcclaywell, @yhoogstrate I can reproduce this, and I think I have fixed it. With the proposed change the above procedure works.
Also @afgane ansible galaxy-tools-playbook is happiliy chugging through more than 50 tools, including multiple revisions, without fail, while before only a minority installed successfully.
@mvdbeek I can verify that this patch eliminates the error I received in the test case above, and that my scripted installation of about 100 tools and dependencies now (mostly) finishes successfully! Thanks for digging this problem up :)
Fixed in https://github.com/galaxyproject/galaxy/pull/681. Thanks @mvdbeek!
I've been working to duplicate all of our installed repositories from production using scripted API calls for reproducibility. I've run into a problem where an installed repository appears to conflict somehow with a later installation of a completely different repository, throwing an unhandled exception and leaving the installing repository in an inconsistent state.
(Note: this is the only failing case I've reduced to the bare minimum from the ~110 tools I was trying to install. Before trying to reduce to a test case I provoked similar errors from at least three other repositories, so the problem doesn't seem to be limited to this combination.)
Against a brand new 15.05 instance, I ran:
followed by:
The first repository, samtools_sort, installed fine. However, installation of htseq_count next failed and logged the following:
This left the htseq_count and package_numpy_1_7 repositories in the "New" state and the package_pysam_0_7_7 repository in the "Cloning" state, even after a restart of the server.
However, without installing samtools_sort first, htseq_count and its dependencies install fine.
from @bcclaywell on trello
(Note from @erasche, untested, moving due to recency + test cases provided)