[ACCUMULO-715] use curator

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Won't Fix
• Component/s: master, tserver
• Assignee: John Vines Assign to me
• Reporter: Eric C. Newton
• Created: 03/Aug/12 18:33
• Updated: 11/Mar/20 21:46
• Resolved: 11/Mar/20 21:46

Description

Curator seems to have distilled a lot of zookeeper lesson's learned. Consider using it as our interface to zookeeper.

Curator advantages:

• handles zookeeper retries, via configurable policies
• works around some very strange failure scenarios
• simplified API
• unlike ZooCache and ZooLock, Curator does not need to be maintained by Accumulo developers

Issue Links

• is related to: ACCUMULO-1572 single node zookeeper failure kills connected accumulo servers
• is related to: ACCUMULO-2348 Consolidate process ZooKeeper Lock[Watcher] code
• is related to: HBASE-10296 Replace ZK with a consensus lib(paxos,zab or raft) running within master processes to provide better master failover performance and state consistency

Sub-Tasks

• Fate's DistributedReadWriteLock shouldn't poll: Open
• Fate's ZooStore should use zoocache or curator: Resolved