Some processes called "background-jobs" in ReSeeD are designed to be run concurrently with, but separately from, the main system. This is to allow processes which may take some time to complete to be run without "blocking" the main system. Examples of such jobs include import/export processes and the DOI-minting process.
Sidekiq
The scheduling and execution of these jobs is managed by a system called Sidekiq. ReSeeD Administrator may access a separate dashboard for the Sidekiq system at the URL:
https://rdms.rd.ruhr-uni-bochum.de/sidkiq
Retrying failed jobs
The Administrator may use the Sidekiq dashboard to identify jobs which have failed - and from here they may restart such jobs if they believe that they may succeed on a second attempt. The screenshot below illustrates this.
DOI-Minting
DOIs are automatically minted for works that have been published. The process for doing this is started in a background job. Occasionally, there may be some fault with the DOI minting process. For example, the DataCite API may not respond in time. Failed DOI Minting jobs can be re-tried through the Sidekiq dashboard (see above).
Recovering from S3 outages
If the S3 system becomes unavailable for some reason, then the ingest jobs will fail. It is possible in such situations for the version of the metadata in S3 to be "behind" the version in Fedora/Postgres.
Once S3 is available again, all that is needed is to re-run these failed jobs: they will always get the latest metadata and persist it to S3.