[
https://issues.apache.org/jira/browse/HBASE-6721?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13742727#comment-13742727
]
Francis Liu commented on HBASE-6721:
------------------------------------
{quote}
The current state of this issue is confusing. The parent is in 'Patch
Available' state but seems stale.
{quote}
The decision was to get the NS patch in and this can follow. Hence it got
stale.
{quote}
Subtask #3 has been committed to trunk/0.96 and 0.94. Other subtasks are
resolved as 'Invalid'. We should at least have a release note describing what
is going to be in 0.96 (and already in 0.94).
{quote}
Subtask #3, is a generic requirement, support for MasterCoprocessor Endpoints.
It's an implementation specific to this jira. And so is the nature of subtask
#6. Do we need release notes for this? I'm not familiar with what needs to get
into release notes. Or how do I update them?
{quote}
What further work is going to be done here? What been abandoned? Has enough
functionality been committed already to address the contributors' use case? Did
something turn out to be problematic?
{quote}
It essentially needs a rebase. Also I'd like to update it to make use of
FavoredNodes. This jira and subtask #4 needs to be completed to get basic
functionality.
{quote}
Nothing more for 0.96 since stack has frozen 0.95, correct? Nothing more need
in 0.96 to support placing tables on certain regionservers only (will help with
0.96 to 0.98 migration story)? I can research the answers to my questions in
commit history but maybe we can get an authoritative answer from Vandana
Ayyalasomayajula or Francis Liu?
{quote}
In terms of backwards compatilibity issues the only piece missing was updating
the LoadBalancer interface which is subtask #6. Tho I did a recheck. The
LoadBalancer interface also needs a configure/initialize method (my bad), see
LoadBalancer in the patch. Hopefully we can get that in 0.96. If not we'll have
to write code to support old LoadBalancers which doesn't have this method.
My memory is a bit rusty but essentially the rest of the code resides in the
custom balancer and the coprocessors. Maybe some internal code changes in AM.
So no backwards compatibility issues apart from the one I mentioned. I'll
double check early next week.
> RegionServer Group based Assignment
> -----------------------------------
>
> Key: HBASE-6721
> URL: https://issues.apache.org/jira/browse/HBASE-6721
> Project: HBase
> Issue Type: New Feature
> Reporter: Francis Liu
> Assignee: Vandana Ayyalasomayajula
> Fix For: 0.96.0
>
> Attachments: 6721-master-webUI.patch, HBASE-6721_10.patch,
> HBASE-6721_8.patch, HBASE-6721_94_2.patch, HBASE-6721_94_3.patch,
> HBASE-6721_94_3.patch, HBASE-6721_94_4.patch, HBASE-6721_94_5.patch,
> HBASE-6721_94_6.patch, HBASE-6721_94_7.patch, HBASE-6721_94.patch,
> HBASE-6721_94.patch, HBASE-6721_9.patch, HBASE-6721_9.patch,
> HBASE-6721-DesigDoc.pdf, HBASE-6721-DesigDoc.pdf, HBASE-6721-DesigDoc.pdf,
> HBASE-6721-DesigDoc.pdf, HBASE-6721_trunk.patch, HBASE-6721_trunk.patch,
> HBASE-6721_trunk.patch
>
>
> In multi-tenant deployments of HBase, it is likely that a RegionServer will
> be serving out regions from a number of different tables owned by various
> client applications. Being able to group a subset of running RegionServers
> and assign specific tables to it, provides a client application a level of
> isolation and resource allocation.
> The proposal essentially is to have an AssignmentManager which is aware of
> RegionServer groups and assigns tables to region servers based on groupings.
> Load balancing will occur on a per group basis as well.
> This is essentially a simplification of the approach taken in HBASE-4120. See
> attached document.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
