[
https://issues.apache.org/jira/browse/HBASE-29203?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ray Mattingly resolved HBASE-29203.
-----------------------------------
Fix Version/s: 4.0.0-alpha-1
2.7.0
3.0.0-beta-2
Release Note: This adds a StoreFileTableSkewCostFunction which will let
you balance every table by datasize, which can be a better metric than region
number. This will work well, even if you are balancing by cluster (the default
balancer setup). It comes with a default cost of 35, equivalent to
StorefileSizeCost
Resolution: Fixed
> There should be a StorefileSize equivalent to the TableSkewCost
> ---------------------------------------------------------------
>
> Key: HBASE-29203
> URL: https://issues.apache.org/jira/browse/HBASE-29203
> Project: HBase
> Issue Type: Improvement
> Components: Balancer
> Affects Versions: 2.7.0
> Reporter: Ray Mattingly
> Assignee: Ray Mattingly
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.0.0-alpha-1, 2.7.0, 3.0.0-beta-2
>
>
> We don't want to rely on per-table balancing, but we do want tables to be
> balanced effectively based on their data size. This is because regions are
> variably sized, but 1gb is always 1gb, so, in a well distributed schema, data
> size is a better metric for a region's load than just counting every region
> identically.
> Right now the TableSkewCostFunction is the balancer's mechanism for balancing
> each table somewhat independently within a cluster-wide plan, but it only
> looks at region counts. We could balance more effectively if we had a
> StoreFileTableSkewCost.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
