[
https://issues.apache.org/jira/browse/HBASE-9976?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13829780#comment-13829780
]
Nicolas Liochon commented on HBASE-9976:
----------------------------------------
bq. We should just drop prefetch. It is of questionable, unproven benefit.
I was thinking about replacing it with a reverse scan (i.e. we do two calls
today, we would do only one), but just removing it might be better. I'm going
to create a jira for this.
But the function "getRowOrBefore" does create a TableName as well, so if we
have 100K regions we will end up with 100K tableName objects, removing the
prefetching will just make the creation slower.
I'm not a big fan of the solution I have here. Just that I don't have a better
one :-)
> Don't create duplicated TableName objects
> -----------------------------------------
>
> Key: HBASE-9976
> URL: https://issues.apache.org/jira/browse/HBASE-9976
> Project: HBase
> Issue Type: Bug
> Components: Client, regionserver
> Affects Versions: 0.98.0, 0.96.0
> Reporter: Nicolas Liochon
> Assignee: Nicolas Liochon
> Fix For: 0.98.0, 0.96.1
>
> Attachments: 9976.v1.patch, 9976.v10.patch, 9976.v11.patch,
> 9976.v4.patch, 9976.v6.patch, 9976.v7.patch, 9976.v7.patch, 9976.v8.patch,
> 9976.v8.patch, 9976.v9.patch
>
>
> A profiling show that the table name is reponsible for 25% of the memory
> needed to keep the region locations. As well, comparisons will be faster if
> two identical table names are a single java object.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
