Hello Drew, Thinking through the delete operation on a strong consistent bucket, a question popped up for me.
Should not we use delete_object instead to keep strong consistency? The delete function takes a key and does not care about the current state/value of the object itself. If in the meantime another process changes the value of an object, delete will simply delete it. So is there a benefit using delete on strongly consistent bucket other than swipe the bucket not caring about other processes store their state info? Thanks for your thoughts, Zsolt -----Eredeti üzenet----- Feladó: riak-users [mailto:riak-users-boun...@lists.basho.com] Meghatalmazó Drew Pirrone-Brusse Küldve: Thursday, January 08, 2015 12:32 AM Címzett: Igor Birman Másolatot kap: riak-users@lists.basho.com Tárgy: Re: Deleting a strongly consistent key Hello Igor, Deletions should be a strongly consistent operation, yes. If after a successful strongly consistent delete a GET of that key returns stale data something is likely wrong, be it with your application, the cluster, or Riak itself. That said, deletions don't actually remove keys from Riak. Strongly consistent deletions should behave identically to conditional PUTs, where the value put is a tombstone. When Riak is running in an eventually consistent mode, those tombstones are (sometimes) eventually harvested and removed from disk. In the strongly consistent mode (as with `delete_mode = keep`) Riak will never harvest those tombstones, so they will remain on disk indefinitely. Our documentation has lots of good information, if you're looking for some implementation details[1]. Generally speaking, if you know you're going to immediately re-use a key, it's better to simply overwrite values than issuing an intermittent delete. Best regards, -Drew [1]: http://docs.basho.com/riak/latest/ops/advanced/strong-consistency/#Implement ation-Details On Wed, Jan 7, 2015 at 10:01 AM, Igor Birman <igor_bir...@yahoo.com> wrote: > Is deleting a strongly consistent key a consistent operation? I have > found in testing that if I delete a key and then attempt to > immediately reuse it, sometimes the delete completes after I have > reused it. Should I expect that when I get a 204 from a delete that I > can immediately create a new key with the same id? > > _______________________________________________ > riak-users mailing list > riak-users@lists.basho.com > http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com > _______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com --- This email has been checked for viruses by Avast antivirus software. http://www.avast.com _______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
