Closed kltm closed 2 months ago
As I collect progress report items, I def think we want this archived--I do not want to be caught needing to regenerate.
Noting that this is a bit of a "non-starter" given that Zenodo doesn't really work anyway with https://github.com/geneontology/pipeline/issues/345
Glarf--the issue may be forced on this release, as recent additions have likely pushed us over the 50GB limit.
51532257933 Apr 5 14:06 go-release-archive.tgz
hm
48G Apr 5 14:06 go-release-archive.tgz
Well, showdown between GB and GiB...
I can confirm we can no longer proceed with current package:
Warning!
Could not upload files.
Uploading the selected files would result in 51.53 GBbut the limit is50.00 GB.
Okay, rebuilding package without these files. I want to look at saving these separately in Zenodo.
Okay, I now have an experimental deposition here: https://zenodo.org/records/10946933, that contains the solr index and the two blazegraph builds. However, it has a "twin" in the dashboard that is marked for "Review". I don't know what that means. I'll check back on it tomorrow. If it looks good:
Zenodo itself seems to be broken at this point--any attempt to make new versions or edit the metadata of the new deposit is met with a vague server error.
Checking again, I seem to have the ability to update the "binaries" package. I think this can be closed out by updating my SOP for the current "semi-manual" releases.
Drop Solr indices from the Zenodo archive bundle, to:
An alternate I want to consider is silently "splitting", maybe asynchronously or after-the-fact, giving us an archive w/o slowing things down.