[ https://issues.apache.org/jira/browse/CASSANDRA-994?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Erick Tryzelaar updated CASSANDRA-994: -------------------------------------- Attachment: (was: 0003-Rename-files-named-EndPoint-to-Endpoint.patch) > re-organize endpointsnitch implementations > ------------------------------------------ > > Key: CASSANDRA-994 > URL: https://issues.apache.org/jira/browse/CASSANDRA-994 > Project: Cassandra > Issue Type: Task > Affects Versions: 0.7 > Reporter: Jonathan Ellis > Assignee: Erick Tryzelaar > Fix For: 0.7 > > > I think the EPS implementations should look like this: > IEndPointSnitch: rename to IEndpointSnitch (lowercase P), otherwise fine as > it is > SimpleSnitch: new default snitch. Implement IEPS directly, with sorting as > no-ops. > AbstractEndpointSnitch: rename to AbstractRackAwareSnitch; replace > isOnSameRack and isInSameDataCenter with getRack and getDatacenter (replacing > getLocation). > EndPointSnitch: rename to RackInferringSnitch > PropertyFileEndpointSnitch: rename to PropertyFileSnitch, move from contrib > to src > DatacenterEndpointSnitch: remove (redundant wrt PFS). "quorum factor per DC" > needs to go in the ReplicationStrategy, not the snitch. > The patch sequence looks like: > 01: perform renames of IEPS, AES, EPS, and PFES > 02: add SS and make method changes to ARAS [was AES] > Separate ticket: DatacenterReplicationStrategy code changes -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.