You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@usergrid.apache.org by "Mike Dunker (JIRA)" <ji...@apache.org> on 2015/11/04 23:25:27 UTC

[jira] [Created] (USERGRID-1098) Perform quorum reads only after waiting and retrying

Mike Dunker created USERGRID-1098:
-------------------------------------

             Summary: Perform quorum reads only after waiting and retrying
                 Key: USERGRID-1098
                 URL: https://issues.apache.org/jira/browse/USERGRID-1098
             Project: Usergrid
          Issue Type: Story
            Reporter: Mike Dunker


If indexing requests make it to another region before C* has replicated the data, and a local quorum read fails, we retry with a quorum read immediately. It would be better to give some time for C* to replicate the data, so we should put the message back in the queue at least once before trying the quorum read.

The delete issues we had show how bad large amounts of quorum reads can be on a C* cluster, so we should avoid where possible.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)