Details
Description
Test starts with a 4 node cluster. (upr enabled)
1) Data is loaded into every node except orchestrator (it only receives replica items).
2) add_stream from orchestrator to other 3 node vbs with data
3) failover and rebalance 2 nodes
4) at this point I expected n_0 should still have streams I added from n_1 plus additional streams it took over
- however memcached crashed.
Port server memcached on node 'babysitter_of_n_0@127.0.0.1' exited with status 139. Restarting. Messages: Wed May 28 21:59:05.832451 EDT 3: (default) UPR (Consumer) eq_uprq:replication:n_1@10.0.0.105->n_0@10.0.0.105:default - (vb 10) Attempting to add stream with start seqno 10, end seqno 18446744073709551615, vbucket uuid 4675300462675623936, snap start seqno 10, and snap end seqno 10
Wed May 28 21:59:05.832735 EDT 3: (default) UPR (Consumer)
repro via cluster_run with 4 nodes and pyupr:
./pyupr -h 10.0.0.105:9000 10.0.0.105:9001 10.0.0.105:9002 10.0.0.105:9003 -b dev -s reb -o test_add_stream_during_failover