Details
Description
Found the issue when running cluster run against watson
1. Create 2 node cluster
2. Stop compaction and expiry pager from running
3. Create default bucket
4. With 10 threads running in parallel,
add items with expiry ranging from 1-100 secs
5. With 10 threads running in parallel,
add items with expiry ranging from 1-100 secs
6. Compaction running
7. Rebalance-in 1 node -> Rebalance-out 1 node -> Rebalance-in 1 node
Steps 5,6, 7 happen in parallel
The second attempt to rebalance-in a node hangs.
Here is the test case
http://review.couchbase.org/55679
./testrunner -i b/resources/dev-6-nodes-xdcr_n1ql_2i.ini -t rebalance.rebalancein.RebalanceInTests.rebalance_in_with_compaction_and_expiration_ops,nodes_init=2,replicas=1,items=10,total_loader_threads=10,expiry_items=100000,max_expiry=100,skip_cleanup=True