Investigate resolver updates

[spacemansteve]

We have seen some incidents where resolver has outdated link data while solr has current/updated data. How is that happening? Is there a bug skips updates to the resolver under certain conditions? If there is an error while updating the resolver will it be retried? Did the testing of the solr index rebuild contribute to this problem? Was it just caused by only updating solr when responding to special requests to process bibcodes (e.g., using -r s rather than -r).

[spacemansteve]

Sergi noted a bug in tasks.py. The code fails to use the correct column name for the datalinks_checksum. So, it fails to appreciate when the resolver data has not changed and always sends an update to resolver.

[spacemansteve]

Sergi noted a bug in app.py. The processed column will be updated even when, for example, the resolver updated failed. This will prevent a failed bibcode from being reset on subsequent runs.