[ 
https://issues.apache.org/jira/browse/HBASE-18946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16290123#comment-16290123
 ] 

stack commented on HBASE-18946:
-------------------------------

TestRestartCluster#testRetainAssignmentOnRestart is broken because when we 
create AssignProcedures now, we populate them w/ target servers. Doing this 
blows our retaining old locations (if no target specified, when AssignProcedure 
runs, it will try to use the new form of the old location -- which is how we 
get our retention). Over in HBASE-19501 we noticed that ServerCrashProcedure 
depends on this and made a version of assign procedures that works for SCP. Let 
me do the same here by reinstituting the old bulk assign method for SCP to use.

This seems to have fixed all tests. Let me post new patch to be sure.

> Stochastic load balancer assigns replica regions to the same RS
> ---------------------------------------------------------------
>
>                 Key: HBASE-18946
>                 URL: https://issues.apache.org/jira/browse/HBASE-18946
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 2.0.0-alpha-3
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>             Fix For: 2.0.0-beta-1
>
>         Attachments: HBASE-18946.master.001.patch, 
> HBASE-18946.master.002.patch, HBASE-18946.master.003.patch, 
> HBASE-18946.master.004.patch, HBASE-18946.master.005.patch, 
> HBASE-18946.master.006.patch, HBASE-18946.patch, HBASE-18946.patch, 
> HBASE-18946_2.patch, HBASE-18946_2.patch, HBASE-18946_simple_7.patch, 
> HBASE-18946_simple_8.patch, TestRegionReplicasWithRestartScenarios.java
>
>
> Trying out region replica and its assignment I can see that some times the 
> default LB Stocahstic load balancer assigns replica regions to the same RS. 
> This happens when we have 3 RS checked in and we have a table with 3 
> replicas. When a RS goes down then the replicas being assigned to same RS is 
> acceptable but the case when we have enough RS to assign this behaviour is 
> undesirable and does not solve the purpose of replicas. 
> [~huaxiang] and [~enis]. 



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to