[
https://issues.apache.org/jira/browse/PHOENIX-2833?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15237898#comment-15237898
]
churro morales commented on PHOENIX-2833:
-----------------------------------------
Forgive me for being crazy.... but the RpcControllerFactory is private and what
we are doing is prioritizing calls to the index tables and phoenix system
tables.
It looks like the HIGH_QOS priority is only used for the meta table and other
important regionserver related operations (like close region, open region,
flush, etc...) which is fine.
Because these are table specific operations for those index and phoenix system
tables; We could simply extend TableName and override
tableName.getSystemTable() to return true which would give a HIGH_QOS priority.
I don't think we need something higher priority than META operations because
if you can't access META that affects all clients and the cluster is in trouble
anyways so given equivalent priority is sufficient.
That way we could get rid of quite a bit of HBase code that depends on private
hbase API's
[~jamestaylor] am I correct in assuming that we don't need a higher priority
than HConstants.HIGH_QOS for an RPC call?
> Upgrade HBase dependency to version 1.2.0
> -----------------------------------------
>
> Key: PHOENIX-2833
> URL: https://issues.apache.org/jira/browse/PHOENIX-2833
> Project: Phoenix
> Issue Type: Task
> Affects Versions: 4.8.0
> Reporter: Rob Leidle
> Priority: Minor
> Fix For: 4.8.0
>
> Attachments: upgrade-to-hbase-1.2.0.patch
>
> Original Estimate: 24h
> Remaining Estimate: 24h
>
> This patch upgrades Phoenix’s HBase dependency version from 1.1 to 1.2. The
> Phoenix version is changed to reflect the HBase change.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
