[
https://issues.apache.org/jira/browse/CASSANDRA-2388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13087915#comment-13087915
]
Patrik Modesto commented on CASSANDRA-2388:
-------------------------------------------
I'd like to point out the situation in which no node for a given range of keys
is available. It can happen for example with keyspace set to RF=1 and a node
goes down. I created a patch that gives a user a chance to ignore missing
range/node and continue runnig the MapReduce job. The patch is here:
http://pastebin.com/hhrr8m9P
Jonathan already replied to the ML with "ignoring unavailable ranges is a
misfeature, imo".
In our case it's very usefull, although there may be another/smarter solution.
We have a keyspace with RF=1 and the nature of our data allows us to ignore
temporarily missing node. The current ColumnFamilyInputFormat fails with
RuntimeException and AFAIK there is no way around.
> 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: 0.7.9
>
> Attachments: 0002_On_TException_try_next_split.patch,
> CASSANDRA-2388-addition1.patch, CASSANDRA-2388-extended.patch,
> CASSANDRA-2388.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
