[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-20 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15508452#comment-15508452 ] Heng Chen commented on HBASE-16652: --- Increment entry size is stable due to it just one long size,

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-20 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15506133#comment-15506133 ] Phil Yang commented on HBASE-16652: --- What are the sizes of two WAL.Entry? > Figure out performance

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15503102#comment-15503102 ] Heng Chen commented on HBASE-16652: --- Will be back later. > Figure out performance difference between

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Duo Zhang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502866#comment-15502866 ] Duo Zhang commented on HBASE-16652: --- This is in {{reckonDeltasByStore}} {code:title=HRegion.java}

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502849#comment-15502849 ] Phil Yang commented on HBASE-16652: --- {code} List currentValues = get(mutation, store, deltas,

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502851#comment-15502851 ] Heng Chen commented on HBASE-16652: --- Let me test it again with cache closed. > Figure out

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502816#comment-15502816 ] Heng Chen commented on HBASE-16652: --- En The time comes from server-side, it is only 'get' action

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502811#comment-15502811 ] Heng Chen commented on HBASE-16652: --- Sounds reasonable. > Figure out performance difference between

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Anoop Sam John (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502798#comment-15502798 ] Anoop Sam John commented on HBASE-16652: Increment is a public exposed class and this will change

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Duo Zhang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502796#comment-15502796 ] Duo Zhang commented on HBASE-16652: --- See ProtobufUtil.toGet and toScan. If TimeRange.isAllTime returns

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502795#comment-15502795 ] Phil Yang commented on HBASE-16652: --- Oh, we have a default TimeRange in Get, too. So the reason that

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502791#comment-15502791 ] Heng Chen commented on HBASE-16652: --- Good question. We could test it, it seems should be faster. >

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502775#comment-15502775 ] Phil Yang commented on HBASE-16652: --- If I am not wrong, we have a default TimeRange in Scan so the tr

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502720#comment-15502720 ] Heng Chen commented on HBASE-16652: --- Found the reason why 'get' in Increment is slower than append.

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-19 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502496#comment-15502496 ] Heng Chen commented on HBASE-16652: --- I add time log about 'get' operation, and do test with pe command,

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-18 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502322#comment-15502322 ] Heng Chen commented on HBASE-16652: --- Thanks [~yangzhe1991]. As profile shows, the main difference

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-18 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502310#comment-15502310 ] Phil Yang commented on HBASE-16652: --- And there may be an improvement that we can add two Longs bit by

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-18 Thread Phil Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502190#comment-15502190 ] Phil Yang commented on HBASE-16652: --- In append we use arraycopy to merge the bytes and in increment we

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-18 Thread Heng Chen (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502137#comment-15502137 ] Heng Chen commented on HBASE-16652: --- Thanks [~allan163]. Increment has one get and one put too

[jira] [Commented] (HBASE-16652) Figure out performance difference between increment and append

2016-09-18 Thread Allan Yang (JIRA)
[ https://issues.apache.org/jira/browse/HBASE-16652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15502128#comment-15502128 ] Allan Yang commented on HBASE-16652: Though may not related to this jira, I'd like to mention that,