Major Description
This is the same issue discussed in CBSE-13981 - rebalancing out a query node whilst a Client/SDK is querying the cluster returns 1181/1180 errors temporarily. Having read through the discussion, I believe the intention was that these errors would no longer be seen by a client during this test against 7.1.4-MP3, with this ticket having been closed MB-56557 and attached to the 7.1.4-MP3 ticket (MB-56555). But they are still occuring reliably in this test case, details below.
Test details:
Server Version: 7.1.4-3632 (7.1.4-MP3)
Cluster start state:
172.23.99.70 - kv
172.23.99.69 - n1ql <- rebalanced out
172.23.96.54 - ndex
172.23.98.17 - n1ql
The test just runs queries throughout from the Java SDK (tested and error observed from other SDKs also), against the cluster, and at the change stage the query node 172.23.99.69 is rebalanced out.
The below graph shows the query results over the course of the test:
http://sdkqe-testresults.couchbase.com.s3.amazonaws.com/SDK-SDK/CB-7.1.4-3632/Rb1OutQuery-N1QL/04-27-23/058907/19500cd6475ab3b8a5215c48ff3c16c9-N1QL.html
At the change phase, a number of 1181/1180 errors occur.
I've attached the cluster logs also.
