as a datapoint, at the time of writing the total (uncompressed) size occupied by content objects that are larger than our current limit is as follows:

softwareheritage=> select count(*), sum(length), avg(length) from content where length > 100 * 1024 * 1024;
 count  |      sum           |        avg         
--------+--------------------+--------------------
 54'803 | 18'605'873'775'395 | 339'504'658

i.e., 50 K contents, for 17 TB total, average size 320 MB.

Depending on the compression level, this might be up to a 5-10% or our current object storage total size.

and here's the actual disk usage (!= compressed size)

/srv/softwareheritage/scratch/lists $ xzcat oversize-contents.txt.xz | while read id ; do du $(swh-ls-obj $id) ; done | cut -f 1 | paste -sd+ | bc
16847900696

those are KB, so the total disk usage is ~15.5 TB (not bad!)

As a byproduct of this, the actual list of objects to move away is available on uffizi, in the file mentioned above.

As a comment on 4, the object_id column is per-table, so you should avoid carrying it over to the skipped_content table.

thanks for the review, I've updated the SQL query accordingly

This is now done on uffizi.
Potentially remaining sub-tasks before closing this:

• removing the oversize objects from banco