[
https://issues.apache.org/jira/browse/HBASE-10357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13977860#comment-13977860
]
Devaraj Das commented on HBASE-10357:
-------------------------------------
[~stack], thanks for taking a look. The patch sure needs cleanup. I was in the
process of doing that. The approach taken here is similar to the one done in
handling GET for cases that have region replicas. In general, the flow is:
1. A scan is attempted on the default region
2. The scanner sends all the RPCs to the default region until it is done, or,
there is a timeout on the default (a timeout of zero should be disallowed). The
primaryOperationTimeout is what is used for this timeout and it is set to a
second by default.
3. If there is a timeout in (2) above, scanner(s) is opened on the non-default
replica(s)
4. The results (and this applies to openScanner as well) from the first
successful scanner are taken, and it is stored which server returned the
results.
5. The next RPCs are done on the above stored server until it is done or there
is a timeout, in which case, the other replicas are queried (as in (3) above).
Does the above give a picture of what is being attempted in the patch? I'll
clean up the patch some and post it shortly.
> Failover RPC's for scans
> ------------------------
>
> Key: HBASE-10357
> URL: https://issues.apache.org/jira/browse/HBASE-10357
> Project: HBase
> Issue Type: Sub-task
> Components: Client
> Reporter: Enis Soztutar
> Fix For: 0.99.0
>
> Attachments: 10357-1.txt, 10357-2.txt, 10357-3.2.txt, 10357-3.txt
>
>
> This is extension of HBASE-10355 to add failover support for scans.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
