[
https://issues.apache.org/jira/browse/HBASE-5974?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13279686#comment-13279686
]
Todd Lipcon commented on HBASE-5974:
------------------------------------
- style: use 2-space indentation, not tabs
- why do we need the new RegionScannerWithCookie class? why not add the cookie
to RegionScanner itself? I'd also move the cookie check into a function
"scanner.checkAndIncrementCookie(cookie);"
- let's rename "cookie" to "callSequenceNumber" or something -- since in this
implementation it is a simple increasing sequence.
- this isn't currently compatible with 0.94, since a new client wouldn't be
able to scan an old server. In order to make it compatible, we'd need to catch
the MethodNotExists exception and set a flag in the client which disables
cookie usage for that cluster.
- In the test, I think you should use HRegionInterface directly, so you don't
have to actually generate an RPC timeout. As is, I think it's also not
guaranteed to trigger the issue unless you set scanner caching to 1, right? I
guess that's the default, but if we ever change the default, it would
invalidate this test.
> Scanner retry behavior with RPC timeout on next() seems incorrect
> -----------------------------------------------------------------
>
> Key: HBASE-5974
> URL: https://issues.apache.org/jira/browse/HBASE-5974
> Project: HBase
> Issue Type: Bug
> Components: client, regionserver
> Affects Versions: 0.90.7, 0.92.1, 0.94.0, 0.96.0
> Reporter: Todd Lipcon
> Priority: Critical
> Attachments: HBASE-5974_0.94.patch
>
>
> I'm seeing the following behavior:
> - set RPC timeout to a short value
> - call next() for some batch of rows, big enough so the client times out
> before the result is returned
> - the HConnectionManager stuff will retry the next() call to the same server.
> At this point, one of two things can happen: 1) the previous next() call will
> still be processing, in which case you get a LeaseException, because it was
> removed from the map during the processing, or 2) the next() call will
> succeed but skip the prior batch of rows.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
