Test publishing with large file

[jcoyne]

Up to 1 TB

Validate the approach implemented in #4965.

[lwrubel]

@andrewjbtw is testing with a large object currently.

[andrewjbtw]

I've created a test object in stage with a 1TB file plus a few smaller files, using data from a real deposit.

To test shelving, it should be sufficient to

1. Take the files off stacks
   (assumes the starting point is "publish=yes,shelve=yes" and files are already on stacks)

• Open a new version
• Change "publish" and "shelve" to "no" to unshelve the files
• Close the version
• Make sure the files are unshelved during accessionWF

2. Put the files on stacks
   (assumes the starting point is "publish=no,shelve=no" and files are not on stacks)

• Open a new version
• Change "publish" and "shelve" to "yes" to shelve the files
• Close the version
• Make sure the files are shelved during accessionWF

[jcoyne]

It looks like https://argo-stage.stanford.edu/view/druid:fb146xw8568 was published, so I think we can say this worked, right?

[andrewjbtw]

I don't think this went through the new API yet. I was just making sure it got into the stage systems for future use.

[jcoyne]

Blocked by sul-dlss/purl-fetcher#751

[jcoyne]

Looks good:

 #<ActiveStorage::Blob:0x00007f2a2e4e0de0
  id: 246,
  key: "ay4ojwqoehp9n9smt5fof1qwoaxy",
  filename: "DTU_DMS.tar.gz",
  content_type: "application/octet-stream",
  metadata: {},
  service_name: "local",
  byte_size: 1196585612576,
  checksum: "6o51iYPNe6/dloBheVB75A==",
  created_at: Fri, 17 May 2024 13:42:04.495049000 UTC +00:00>