Method and apparatus for eventually consistent delete in a distributed data store

摘要

Techniques for effective delete operations in a distributed data store with eventually consistent replicated entries include determining to delete a particular entry from the distributed data store. Each entry includes a first field that holds data that indicates a key and a second field that holds data that indicates content associated with the key and a third field that holds data that indicates a version for the content. The method also comprises causing, at least in part, actions that result in marking the particular entry as deleted without removing the particular entry, and updating a version in the third field for the particular entry.

主权项

1. A method comprising:
determining to delete a particular entry from a distributed data store with eventually consistent replicated entries, wherein each entry includes a first field that holds a key, a second field that holds content associated with the key, and a third field that holds a version for the content; causing, at least in part, a marking of the particular entry as deleted without removing the particular entry from the distributed data store in response to the determination to delete the particular entry; causing, at least in part, an updating of the version in the third field for the particular entry in response to the marking of the particular entry; determining whether all returned replicates of the particular entry include identical value in the corresponding second fields in response to a get command for the particular entry; and if all of the returned replicates of the particular entry do not include the identical value in the corresponding second fields, determining a correct value for each of the corresponding second fields, for eventually consistent delete in the distributed data store.