[
https://issues.apache.org/jira/browse/HBASE-29708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Istvan Toth updated HBASE-29708:
--------------------------------
Description:
Some Phoenix tests aggressively moving regions during scans are failing with
HBase 3.
-Seems to be caused by AsyncScanSingleRegionRpcRetryingCaller ignoring
heartbeat responses when setting includeNextStartRowWhenError.-
This is caused by phoenix "dummy cell" mechanism, which confuses the HBase side
scan retry logic.
was:
Some Phoenix tests aggressively moving regions during scans are failing with
HBase 3.
Seems to be caused by AsyncScanSingleRegionRpcRetryingCaller ignoring heartbeat
responses when setting includeNextStartRowWhenError.
> Heartbeat messages can cause scans to be continued from the wrong row
> ---------------------------------------------------------------------
>
> Key: HBASE-29708
> URL: https://issues.apache.org/jira/browse/HBASE-29708
> Project: HBase
> Issue Type: Bug
> Affects Versions: 3.0.0-beta-2
> Reporter: Istvan Toth
> Assignee: Istvan Toth
> Priority: Major
> Labels: pull-request-available
>
> Some Phoenix tests aggressively moving regions during scans are failing with
> HBase 3.
> -Seems to be caused by AsyncScanSingleRegionRpcRetryingCaller ignoring
> heartbeat responses when setting includeNextStartRowWhenError.-
> This is caused by phoenix "dummy cell" mechanism, which confuses the HBase
> side scan retry logic.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
