Details
Description
Install Couchbase Server 4.6.2-3900 on 8 Centos 6.x servers
Create an XDCR with replication from cluster A (41, 83, 84, 58 and 59) to cluster B (19, 180 and 182)
Create default bucket on cluster A.
Load 10K in cluster A. Replication goes to cluster B
At cluster A, do hard failover 2 nodes (141 and 83).
Run cbrecovery to recover data from cluster B to cluster A
/opt/couchbase/bin/cbrecovery http://172.23.107.19:8091 http://172.23.106.84:8091 -b default -B default -u Administrator -p password -U Administrator -P password |
The test failed due to the failure of recovery 2 last vbuckets [852, 853] and hang forever.
Result is cluster B shows 10K items but cluster A only shows 9992 items
Here is the command to run using testrunner.
python testrunner.py -i ../../ini/8-c-tools.ini -t cbRecoverytests.cbrecovery.cbrecover_multiple_failover_swapout_reb_routine,items=10000,rdirection=unidirection,ctopology=chain,failover=source,fail_count=2,add_count=2,max_verify=10000 |
|
|
I will check the last build this test passed and update here.