Details
-
Bug
-
Resolution: Fixed
-
Major
-
5.5.4
-
None
-
5.5.4-4340-enterprise
Description
Script to Repro
./testrunner -i /tmp/testexec.22233.ini -p get-cbcollect-info=True,bucket_type=ephemeral -t rebalance.rebalance_high_ops_pillowfight.RebalanceHighOpsWithPillowFight.test_graceful_failover_addback,node_out=1,replicas=1,nodes_init=1,items=2000000,batch_size=1000,rate_limit=300000,recovery_type=full,instances=2,threads=5,loader=pillowfight,flusher_batch_split_trigger=1
|
Steps to Repro
1. Created a =1 node cluster with 1 replica and create default bucket.
2.Set ns_bucket:update_bucket_props("default", [
]). and restart the memcached.
3. Start data load
4. While 3 in progress add 1 node and rebalance in.
5. Start data load.
6. Now do a graceful failover of newly added node, do full recovery and cancel add back. Do this a few times.
7. Do a graceful failover. Do full recovery and rebalance.
Rebalance fails as shown below.
{u'node': u'ns_1@172.23.105.103', u'code': 0, u'text': u'Rebalance exited with reason {mover_crashed,\n {unexpected_exit,\n {\'EXIT\',<0.6530.42>,\n {dcp_wait_for_data_move_failed,"default",3,\n \'ns_1@172.23.105.103\',\n [\'ns_1@172.23.105.105\'],\n {error,no_stats_for_this_vbucket}}}}}', u'shortText': u'message', u'serverTime': u'2019-05-28T03:13:16.400Z', u'module': u'ns_orchestrator', u'tstamp': 1559038396400, u'type': u'critical'}
|
See cbcollect_info attached.