Details
-
Bug
-
Resolution: Fixed
-
Critical
-
6.5.0
-
Untriaged
-
Unknown
Description
1. Add 60 second sleep in recoverableCreateIndex after plan phase, before the commit phase.
2. Create a cluster run setup with 4 nodes. Create default bucket.
Node1:kv+query+index
Node2:query+index
Node3:query+index
Node4:query+index
3. On query node 127.0.0.1:9000 run create index command:
create index index1 on default(a) with {"nodes": ["127.0.0.1:9001"]} |
4. Within next 2/3 seconds, on query node 127.0.0.1:9001 run create index command:
create index index1 on default(b) with {"nodes": ["127.0.0.1:9002"]} |
5. Wait for 60 seconds. Here, both the create index statements succeed. Two different indexes (with same name) can be seen on UI on two different nodes.
Attachments
Issue Links
- relates to
-
DOC-4321 CREATE INDEX - default index placement policy needs to be documented
- Resolved