> +A Job represents a request we send to Glacier to read data. When a job > finishes it produces an output > +that we can read. The estimated time to finish a job is ~4 hours. > + > +There are two different kinds of jobs: > +* Inventory retrieval: Used to list the archives within a vault. > +* Archive retrieval: Used to read an archive. > + > +**Note**: Inventories are updated only once every 24 hours. The archive list > in an inventory may be outdated. > + > +There are also two different ways to discover if a job has finished: by > polling or setting a Notification-Configuration to the vault. This > Notification-Configuration will notify Amazon SNS when is the data ready to > be retrieved. Only one Notification-Configuration can be set per Vault. > + > +At the time of writing only polling is supported by the jclouds Glacier > provider. > + > +BlobStore View > +-------------- > +Due to the Glacier's nature, many of the BlobStore view methods are will be > very slow. Many Glacier users maintain their own application index of Glacier > archives to this. In addition, you can interrupt long-running Glacier > operations by sending an interrupt to the thread and catching the resulting > exception.
> Just to check I'm understanding this correctly...you mean an index of Glacier > archives in their application logic? Correct, many applications cache this information client-side. In fact, @rcoedo 's original implementation used this approach but we removed it so that we can get feedback from users before committing to a particular approach. --- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds-site/pull/121/files#r16034904
