Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.

Once you have successfully ingested data to ECHO that data must go through the following steps to be visible.

Synchronous processes

  1. Persistence of metadata to ECHO
  2. Persistence of metadata to CMR

Asynchronous processes

  1. Indexing within ECHO
  2. Indexing within CMR

Once the above have been achieved any dataset or granule data is visible subject to ACL restrictions from our catalog-rest API.

In operations, indexing usually occurs within seconds of persistence.

The Catalog-Rest cache

Catalog-rest caches access control list (ACL) information. This is refreshed every 15 minutes. If you modify an ACL to expose inventory then this inventory will not be visible for up to 15 minutes.

The Reverb cache

Reverb caches dataset metadata (including granule counts associated with a dataset). This cache is refreshed on the hour. This means the following caveats associated with visibility of inventory in Reverb.

  1. New datasets may take up to an hour to be visible in Reverb.
  2. Updated dataset information may take up to an hour to be visible in Reverb.
  3. Granules that have been added to a new dataset or a dataset that had no granules previously may take up to an hour to be searchable in Reverb (Reverb caches granule counts associated with a dataset every hour so if at hour 1 it had no granules and you add some after that the search capability will not be enabled until the cache refreshes on hour 2)

Other Reverb cache considerations

Reverb caches the following information every hour

  1. Dataset metadata
  2. Dataset holdings - (this gives rise to the phenomena of users not seeing granules for datasets who's holdings were previously zero, this often gives the providers the impression that granules are cached in Reverb. They are not)
  3. Service information relating to datasets and granules
  4. Whether datasets can be put in the shopping cart
  5. Calendar events


Hide comments