[jira] [Resolved] (HBASE-15678) Normalize RetryingCallable cache clearing and implementations

["Andrew Purtell (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-15678?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Andrew Purtell resolved HBASE-15678.
------------------------------------
       Resolution: Incomplete
         Assignee:     (was: Gary Helmling)
    Fix Version/s:     (was: 1.3.3)

> Normalize RetryingCallable cache clearing and implementations
> -------------------------------------------------------------
>
>                 Key: HBASE-15678
>                 URL: https://issues.apache.org/jira/browse/HBASE-15678
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Client
>            Reporter: Gary Helmling
>            Priority: Major
>
> This is a fair amount of duplication and inconsistency in the meta cache handling of RetryingCallable implementations:
> * meta cache is often cleared in prepare() when reload=true, in addition to being cleared in throwable()
> * each RetryingCallable implementation does this slightly differently, leading to inconsistencies and potential bugs
> * RegionServerCallable and RegionAdminServiceCallable duplicate a lot of code, but with small, seemingly unnecessary inconsistencies.  We should clean these up into a common base with subclasses doing only the necessary differentiation.
> The main goal here is to establish some common handling, to the extent possible, for the meta cache interactions by the different implementations.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)