Page History

...

1. node_i sees invalid files in ACE_i
2. node_i gathers invalid files and issues a repair request to the ingest server
   
   1. POST /api/repair
   2. Handled manually
   3. Consider having multiple requests in the event many files are corrupt
3. node_v sees the repair request
   1. Handled manually, likely from discussion in the chron group
4. node_v checks ACE_v to see if the files are valid
   1. POST /api/repair/<id>/fulfill if valid
5. node_v stages content for node_i
   
   1. P2P: make a link (or links) to the files in a directory for node_i
      
   2. Ingest: rsync the files up to the ingest server
   3. ACE: create a token for node_i and make that available
6. node_v notifies ingest server that content is ready for node_i
   1. POST /api/repair/fulfillment/<id>/ready
7. node_i replicates staged content
   1. GET /api/repair/fulfillment?to=node_i&status=ready
8. node_i validates staged content
   1. communicates with ACE compare API
   2. if not valid, end here
9. node_i copies staged content to preservation storage
10. node_i issues an audit of the corrupt files
11. node_i responds with the result of the audit
    1. if the audit is not successful a new replication request will need to be made, but we might want to do that by hand
    2. POST /api/repair/fulfillment/<id>/complete

...