[ https://issues.apache.org/jira/browse/CASSANDRA-2388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13056546#comment-13056546 ]
Jonathan Ellis commented on CASSANDRA-2388: ------------------------------------------- bq. CFRR's local node is the right and only node worth sorting against, it being the "task tracker node". Right. bq. Then it is a random c* node that becomes the "local node" We still want to sort by proxmity-to-TT, because CFRR connects directly to the split owner to do the reads. initialAddress isn't involved post-split-discovery. Again, all the complexity goes away if we just embed the snitch into CFIF/TT. One wrinkle: ec2snitch requires gossip, so TT would need a separate local ip to participate in the gossip ring. We could make that optional (and fall back to old "recognize local data, otherwise you get a random replica" behavior otherwise). > ColumnFamilyRecordReader fails for a given split because a host is down, even > if records could reasonably be read from other replica. > ------------------------------------------------------------------------------------------------------------------------------------- > > Key: CASSANDRA-2388 > URL: https://issues.apache.org/jira/browse/CASSANDRA-2388 > Project: Cassandra > Issue Type: Bug > Components: Hadoop > Affects Versions: 0.7.6, 0.8.0 > Reporter: Eldon Stegall > Assignee: Jeremy Hanna > Labels: hadoop, inputformat > Fix For: 0.7.7, 0.8.2 > > Attachments: 0002_On_TException_try_next_split.patch, > CASSANDRA-2388.patch, CASSANDRA-2388.patch, CASSANDRA-2388.patch > > > ColumnFamilyRecordReader only tries the first location for a given split. We > should try multiple locations for a given split. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira