Details
-
Improvement
-
Status: Closed
-
Minor
-
Resolution: Invalid
-
HBASE-12259
-
None
-
None
Description
I'm very interested in the Raft implementation (in particular, whether it can be reused elsewhere, given that this effort seems likely to produce the canonical Raft implementation).
I'd like to try it standalone, but I'm not yet having much success.
I start a cluster of 3 servers (N=0,1,2):
java -cp "conf:hbase-consensus/target/hbase-consensus-2.0.0-SNAPSHOT.jar:hbase-consensus/target/dependency/*" org.apache.hadoop.hbase.consensus.server.LocalConsensusServer -region 1 -servers 127.0.0.1:10000,127.0.0.1:10001,127.0.0.1:10002 -debug DEBUG -localIndex ${N}
And then I try running the load-test client:
java -cp "hbase-consensus/target/hbase-consensus-2.0.0-SNAPSHOT.jar:hbase-consensus/target/dependency/*" org.apache.hadoop.hbase.consensus.client.QuorumLoadTestClient -region 1 -servers 127.0.0.1:10000,127.0.0.1:10001,127.0.0.1:10002
That gives an immediate NPE at org.apache.hadoop.hbase.consensus.quorum.QuorumInfo.populateInternalMaps(QuorumInfo.java:315).
Are there any docs on how to get started playing with running this branch? Thanks!