[ 
https://issues.apache.org/jira/browse/HBASE-29252?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Charles Connell updated HBASE-29252:
------------------------------------
    Description: 
I've looked at a lot of allocation profiles of RegionServers doing a read-heavy 
workload. Some allocations that dominate the chart can be easily avoided. The 
class {{RowIndexSeekerV1.SeekerState}} contains a {{ByteBufferKeyOnlyKeyValue}} 
field that is replaced on every row read. This object can be reset and re-used 
instead.

On the attached profile, allocations of this object account for 9% of the 
allocations done. This is from a load test and is not necessarily 
representative of your traffic patterns.

  was:I've looked at a lot of allocation profiles of RegionServers doing a 
read-heavy workload. Some allocations that dominate the chart can be easily 
avoided. The class {{RowIndexSeekerV1.SeekerState}} contains a 
{{ByteBufferKeyOnlyKeyValue}} field that is replaced on every row read. On the 
attached profile, this accounts for 9% of the allocations done. This object can 
be reset and re-used instead.


> Avoid allocating a new ByteBufferKeyOnlyKeyValue on every row processed by 
> RowIndexSeekerV1
> -------------------------------------------------------------------------------------------
>
>                 Key: HBASE-29252
>                 URL: https://issues.apache.org/jira/browse/HBASE-29252
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Charles Connell
>            Assignee: Charles Connell
>            Priority: Minor
>         Attachments: scenario-alloc-hs26.html
>
>
> I've looked at a lot of allocation profiles of RegionServers doing a 
> read-heavy workload. Some allocations that dominate the chart can be easily 
> avoided. The class {{RowIndexSeekerV1.SeekerState}} contains a 
> {{ByteBufferKeyOnlyKeyValue}} field that is replaced on every row read. This 
> object can be reset and re-used instead.
> On the attached profile, allocations of this object account for 9% of the 
> allocations done. This is from a load test and is not necessarily 
> representative of your traffic patterns.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to