Details
-
Bug
-
Resolution: Fixed
-
Critical
-
3.0
-
Security Level: Public
-
None
-
Untriaged
-
Unknown
Description
Run:
NODES=4 TEST=rebalance.rebalanceinout.RebalanceInOutTests.measure_time_index_during_rebalance,items=200000,data_perc_add=30,nodes_init=3,nodes_in=1,skip_cleanup=True,nodes_out=1,num_ddocs=2,num_views=2,max_verify=50000,value_size=1024,GROUP=IN_OUT make any-test
Index group gets reset and query fails while rebalance. The rootcause of index reset is view engine trying to rollback to seq 0 for all vbuckets and it resets the index file.
There is some corruption happening around updation of partition versions structure at some point during the view engine operation which is causing invalid rollback.
Attachments
Issue Links
- relates to
-
MB-11672 Missing items in index after rebalance (Intermittent failure)
- Closed