[
https://issues.apache.org/jira/browse/HBASE-15811?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15279572#comment-15279572
]
stack commented on HBASE-15811:
-------------------------------
This is a bit ugly. Its mvcc. If I immediately go back again when I get an
empty result, it works. If I log the mvcc between the get that fails and the
one that succeeds, we've moved on by a single point. Here is my hack patch:
{code}
diff --git
a/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
index e620c60..93319c3 100644
---
a/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
+++
b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
@@ -6868,13 +6868,24 @@ public class HRegion implements HeapSize,
PropagatingConfigurationObserver, Regi
long before = EnvironmentEdgeManager.currentTime();
Scan scan = new Scan(get);
- RegionScanner scanner = null;
- try {
- scanner = getScanner(scan);
- scanner.next(results);
- } finally {
- if (scanner != null)
- scanner.close();
+ long lastmvcc = -1l;
+ while (true) {
+ RegionScanner scanner = null;
+ try {
+ scanner = getScanner(scan);
+ scanner.next(results);
+ if (!results.isEmpty() && results.get(0).getRow() != null) {
+ if (lastmvcc != -1l) {
+ LOG.info("SPIN EXIT CURRENT MVCC " + scanner.getMvccReadPoint() +
"; last mvcc=" + lastmvcc);
+ }
+ break;
+ }
+ lastmvcc = scanner.getMvccReadPoint();
+ LOG.info("SPIN EMPTY " + lastmvcc);
+ } finally {
+ if (scanner != null)
+ scanner.close();
+ }
}
{code}
Here is the output:
{code}
10 2016-05-10 22:25:18,283 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EMPTY 662
11 2016-05-10 22:25:18,284 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EXIT CURRENT MVCC 663; last mvcc=662
12 2016-05-10 22:25:18,286 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EMPTY 666
13 2016-05-10 22:25:18,286 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EXIT CURRENT MVCC 667; last mvcc=666
14 2016-05-10 22:25:18,287 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EMPTY 666
15 2016-05-10 22:25:18,288 INFO
[B.defaultRpcServer.handler=45,queue=45,port=16020] regionserver.HRegion: SPIN
EXIT CURRENT MVCC 667; last mvcc=666
{code}
... in other words, after getting new Scanner, we've moved the read point on by
one so we see the Put that just happened.
> Batch Get after batch Put does not fetch all Cells
> --------------------------------------------------
>
> Key: HBASE-15811
> URL: https://issues.apache.org/jira/browse/HBASE-15811
> Project: HBase
> Issue Type: Bug
> Components: Client
> Affects Versions: 1.2.1
> Reporter: stack
> Assignee: stack
> Attachments: Test.java
>
>
> A big batch put followed by a batch get does not always return all Cells put.
> See attached test program by Robert Farr that reproduces the issue. It seems
> to be an issue to do with a cluster of more than one machine. Running against
> a single machine does not have the problem (though the single machine may
> have many regions). Robert was unable to make his program fail with a single
> machine only.
> I reproduced what Robert was seeing running his program. I was also unable to
> make a single machine fail. In a batch of 1000 puts, I see one to three Gets
> fail. I noticed too that if I wait a second after a fail and then re-get, the
> Get succeeds.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
