[ https://issues.apache.org/jira/browse/HBASE-20700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16505792#comment-16505792 ]
Duo Zhang commented on HBASE-20700: ----------------------------------- OK seems letting meta online first can make life much easier... So I think we could do the following: 1. Introduce a new type of queue in MasterProcedureScheduler called meta, only RecoverMetaProcedure can be put into this queue. 2. Introduce a new type of lock for RMP, do not use table lock any more so that RMP will not be blocked by MRP. 3. MasterProcedureScheduler will be started with a flag which only enable polling procedures out from meta queue, and after the meta region is online, we disable the flag and then everything back to normal. > Move meta region when server crash can cause the procedure to be stuck > ---------------------------------------------------------------------- > > Key: HBASE-20700 > URL: https://issues.apache.org/jira/browse/HBASE-20700 > Project: HBase > Issue Type: Sub-task > Components: master, proc-v2, Region Assignment > Reporter: Duo Zhang > Assignee: Duo Zhang > Priority: Critical > Fix For: 3.0.0, 2.1.0, 2.0.1 > > Attachments: HBASE-20700-UT.patch > > > As said in HBASE-20682. -- This message was sent by Atlassian JIRA (v7.6.3#76005)