[
https://issues.apache.org/jira/browse/HBASE-19397?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Duo Zhang updated HBASE-19397:
------------------------------
Release Note:
Introduce 5 procedures to do peer modifications:
AddPeerProcedure
RemovePeerProcedure
UpdatePeerConfigProcedure
EnablePeerProcedure
DisablePeerProcedure
The procedures are all executed with the following stage:
1. Call pre CP hook, if an exception is thrown then give up
2. Check whether the operation is valid, if not then give up
3. Update peer storage. Notice that if we have entered this stage, then we can
not rollback any more.
4. Schedule sub procedures to refresh the peer config on every RS.
5. Do post cleanup if any.
6. Call post CP hook. The exception thrown will be ignore since we have already
done the work.
The procedure will hold an exclusive lock on the peer id, so now there is no
concurrent modifications on a single peer.
And now we can guarantee that once the procedure is done, the peer modification
has already taken effect on all RSes.
We have also abstracted a storage layer for replication peer/queue manangement,
and refactored the upper layer to remove zk related naming/code/comment.
> Design procedures for ReplicationManager to notify peer change event from
> master
> ---------------------------------------------------------------------------------
>
> Key: HBASE-19397
> URL: https://issues.apache.org/jira/browse/HBASE-19397
> Project: HBase
> Issue Type: New Feature
> Components: proc-v2, Replication
> Reporter: Zheng Hu
> Assignee: Zheng Hu
>
> After we store peer states / peer queues information into hbase table, RS
> can not track peer config change by adding watcher znode.
> So we need design procedures for ReplicationManager to notify peer change
> event. the replication rpc interfaces which may be implemented by
> procedures are following:
> {code}
> 1. addReplicationPeer
> 2. removeReplicationPeer
> 3. enableReplicationPeer
> 4. disableReplicationPeer
> 5. updateReplicationPeerConfig
> {code}
> BTW, our RS states will still be store in zookeeper, so when RS crash, the
> tracker which will trigger to transfer queues of crashed RS will still be a
> Zookeeper Tracker. we need NOT implement that by procedures.
> As we will release 2.0 in next weeks, and the HBASE-15867 can not be
> resolved before the release, so I'd prefer to create a new feature branch
> for HBASE-15867.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
