Details
-
Bug
-
Resolution: Unresolved
-
Minor
-
5.5.0
-
5.5.0-2873-enterprise
-
Untriaged
-
Centos 64-bit
-
-
Unknown
Description
Script to Repro
./testrunner -i /tmp/upgrade1.ini -p get-cbcollect-info=True -t 2i.gsi_rebalance_2i.SecondaryIndexingRebalanceTests.test_kv_and_gsi_rebalance_with_high_ops,doc-per-day=1000,dataset=array,nodes_init=6,services_init=kv-kv-kv-index-index-n1ql,groups=simple,reset_services=True,skip_cleanup=True,kv_node_out=1,index_node_out=3,num_docs=7000000,skip_cleanup=True
|
This script basically does KV and GSI swap rebalance at high ops rate(around 70k ops/sec). What I noticed is rebalance increases, decreases and then increases again. This happened a couple of times. This could be pretty confusing to the end customer.
Logs attached.
Logs from automation run that confirms this - index_rebalance_percentage