[
https://issues.apache.org/jira/browse/HBASE-15267?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15148388#comment-15148388
]
Heng Chen commented on HBASE-15267:
-----------------------------------
I think i found the reason why scan ingore inexistent row, but read got it.
As current logic, readRequestCount update by scan is different with get.
Relates code like below:
{code: title=HRegion.java}
2731 if (!values.isEmpty()) {
final boolean partial =
scannerContext.partialResultFormed();
Result r = Result.create(values, null, stale, partial);
lastBlock = addSize(context, r, lastBlock);
results.add(r);
i++;
}
.........
2786 region.updateReadRequestsCount(i);
{code}
And "i" updated only when scan result is not empty.
So it could explain why scan ignore the inexistent row.
> Read requests count metric is increased when Get operation does not return
> any record
> -------------------------------------------------------------------------------------
>
> Key: HBASE-15267
> URL: https://issues.apache.org/jira/browse/HBASE-15267
> Project: HBase
> Issue Type: Bug
> Reporter: Eungsop Yoo
> Priority: Minor
> Attachments: HBASE-15267-v1.patch, HBASE-15267.patch
>
>
> Read requests count is increased when Get operation returns no record by
> filtering out. In such cases, Get for deleted record, Get for TTL expired
> record and Get for filtering out.
> So I fixed the bug and added some more test cases for read metrics.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
