[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15712352#comment-15712352
]
Hudson commented on HBASE-16648:
FAILURE: Integrated in Jenkins build HBase-Trunk_matrix #2053 (See
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15711806#comment-15711806
]
Hadoop QA commented on HBASE-16648:
---
| (/) *{color:green}+1 overall{color}* |
\\
\\
|| Vote ||
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15708529#comment-15708529
]
Hadoop QA commented on HBASE-16648:
---
| (/) *{color:green}+1 overall{color}* |
\\
\\
|| Vote ||
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15707765#comment-15707765
]
Hadoop QA commented on HBASE-16648:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15508510#comment-15508510
]
Duo Zhang commented on HBASE-16648:
---
OK, I found the problem...
The map will not be empty every time
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15508439#comment-15508439
]
Duo Zhang commented on HBASE-16648:
---
Seems the critical operation is the get before putIfAbsent, we do
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15508357#comment-15508357
]
Hiroshi Ikeda commented on HBASE-16648:
---
That result seems natural. The atomicity of
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15506758#comment-15506758
]
Duo Zhang commented on HBASE-16648:
---
I'm not Doug Lee so it is hard to tell what exactly happens here.
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15506690#comment-15506690
]
Duo Zhang commented on HBASE-16648:
---
Rewrite the JMH test and run it again.
{noformat}
./bin/run.sh -f
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15506090#comment-15506090
]
Duo Zhang commented on HBASE-16648:
---
Maybe I should use a new CHM instance for each iteration?
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15506063#comment-15506063
]
Hiroshi Ikeda commented on HBASE-16648:
---
I'm not sure about JMH, but I suspect that the classes are
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505739#comment-15505739
]
Duo Zhang commented on HBASE-16648:
---
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505666#comment-15505666
]
Duo Zhang commented on HBASE-16648:
---
Write a simple JMH test, it shows that the default implementation
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505655#comment-15505655
]
Hiroshi Ikeda commented on HBASE-16648:
---
My bad, I didn't realized that the default implementation
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505595#comment-15505595
]
Hadoop QA commented on HBASE-16648:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505360#comment-15505360
]
Duo Zhang commented on HBASE-16648:
---
Lock does not reduce performance, the problem is contention. And I
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15505335#comment-15505335
]
Hiroshi Ikeda commented on HBASE-16648:
---
I read just halfway the patch, but behaviors of
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15503423#comment-15503423
]
Duo Zhang commented on HBASE-16648:
---
[~enis] This is also a benefit from JDK8 where we could use
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15501226#comment-15501226
]
Hadoop QA commented on HBASE-16648:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15501015#comment-15501015
]
Duo Zhang commented on HBASE-16648:
---
The behavior of reportServerError method in
[
https://issues.apache.org/jira/browse/HBASE-16648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15500736#comment-15500736
]
Hadoop QA commented on HBASE-16648:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
21 matches
Mail list logo