search

galaxyproject / idc

Simon's Data Club - Reference data for Galaxy servers
MIT License
9 stars 7 forks source link

Kraken2 reference index mixup #37

Open jennaj opened 4 months ago

jennaj commented 4 months ago

One of the Kraken2 indexes is an exact duplicate of another, instead of containing the actual distinct data.

So -- this needs a new index to be created (or found?), and replaced over the existing incorrect file. From what I can tell, impacts all usegalaxy servers and any others using the CVMFS mount.

If I can help with this, I will, but I'm not sure where the originals are being created. I should already have access to CVMFS and probably can figure out the transfer details etc.

Compare, these are nearly the same. Missing plants?

http://datacache.galaxyproject.org/managed/kraken2_databases/2022-09-05T092205Z_standard_prebuilt_pluspfp_2022-06-07/inspect.txt http://datacache.galaxyproject.org/managed/kraken2_databases/2022-09-04T165121Z_standard_prebuilt_pluspf_2022-06-07/inspect.txt Ran some comparisons in this history. Differences are minor, and look more like database updates rather than an entirely new kingdom added/missing. https://usegalaxy.org/u/jen-galaxyproject/h/kraken2-indexes

Seems to have been introduced upstream, not just in CVMFS, since presents at EU as well. Ticket at EU:

Reported at https://help.galaxyproject.org/t/kraken2-databases-mixed-up/11290/6

Reported at

Consolidated from

vebaev commented 3 months ago

Maybe the issue is not Galaxy related, but within the original pre-compiled indexes please see: https://github.com/DerrickWood/kraken2/issues/785#issuecomment-2009838656

vebaev commented 1 week ago

The DBs have update this month so hopefully the admins can update Galaxy DBs and the issue will be fixed soon. ...ping @bgruening :)