This is a Fault Tolerance Experiment using Raft Step 1: Run the client.py program as:
python2.7 client.py 6002 6003 6004 6005 6006 in on window python2.7 client.py 6003 6002 6004 6005 6006 in on window python2.7 client.py 6004 6003 6002 6005 6006 in on window python2.7 client.py 6005 6003 6002 6004 6006 in on window python2.7 client.py 6006 6003 6002 6005 6004 in on window
On the client input number: 6 pop 7 create 8 add 9 print 10 show id 11 top 12 size
To kill a node simply "cmd c" or grep kill -p on a window. This terminates a node.
syncobj_admin -conn 127.0.0.1:6002 -status shows the status and leader of the cluster.
syncobj_admin -conn 127.0.0.1:6002 -add 127.0.0.1:6010 adds a cluster.
The program does: Leader election Dynamic membership change Log replication Encryption Tolerates 2 terminations Log compactions
Doesn't do:
Does no detect the least amount of nodes for the raft consensus to work. No unified client