[
https://issues.apache.org/jira/browse/PHOENIX-4936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16765572#comment-16765572
]
Xinyi Yan commented on PHOENIX-4936:
------------------------------------
It has empty/bad results problem with HBase 0.98, but not with HBase 1.1+. This
is because HBase 0.98 has no way to enforce a timeout limit on a scan RPC
request. For a large table, such as SYSTEM.CATALOG, the region server takes a
while to execute the scan, and it might cause:
# timeouts on the client side.
# an empty/bad result is returned by the region server.
The region server cuts time and returns whatever results it has accumulated up
to that point to the client, and the client side can't interpret this as an
exhausted table/region or no results are retrieved.
Empty/bad result issues can be found at HBase Jira
https://issues.apache.org/jira/browse/HBASE-13090
Should we close this issue? [~vincentpoon] [~tdsilva] [~karanmehta93]
> Empty resultset returned when hbase.rpc.timeout hit
> ---------------------------------------------------
>
> Key: PHOENIX-4936
> URL: https://issues.apache.org/jira/browse/PHOENIX-4936
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.14.0
> Reporter: Vincent Poon
> Assignee: Xinyi Yan
> Priority: Blocker
> Labels: SFDC
> Fix For: 4.15.0, 5.1.0
>
>
> Seeing this on a large syscat table (~11gb)
> From sqlline, issue a SELECT statement which does a full table scan
> hbase.rpc.timeout gets hit, and instead of getting an exception, an empty
> resultset is silently returned.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
