Details
-
Bug
-
Resolution: Fixed
-
Critical
-
4.0.0
-
Security Level: Public
-
None
-
dev env: 04/09 12:00 pm rep sync build
-
Untriaged
-
Unknown
Description
./testrunner -i b/resources/dev-6-nodes-xdcr_n1ql_2i.ini -t 2i.recovery_2i.SecondaryIndexingRecoveryTests.test_rebalance_in,initial=,before=create_index,after=query_ops,groups=simple,dataset=default,doc-per-day=50,nodes_init=4,nodes_in=2,services_in=kv-kv,services_init=n1ql:kv-kv-index-index,skip_cleanup=True,run_drop_index=False,timeout_for_index_online=600
1. Create cluster with 4 nodes (1: kv,n1ql 2: kv 3:index 4:index
2. Create default bucket with 100K items - employee data set
3. Create primary index using views, wait till it comes online
4. Create 10 GSI indexes in sequence, wait till they come online
5. Create 2 GSI indexes and build in parallel
CREATE INDEX employeecb82125e3c0142a288d4747c2e6a839bjob_title ON default(job_title) USING GSI
CREATE INDEX employeecb82125e3c0142a288d4747c2e6a839bjoin_yr ON default(join_yr) USING GSI
6. Rebalance-in 2 kv nodes
Operation 5 and 6 happen in parallel. The indexes do not come online even after 10 minutes.
Attachments
For Gerrit Dashboard: MB-14357 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
49883,11 | MB-14357 : rebalance 2 kv nodes hangs due to race condition of StreamBegin | unstable | indexing | Status: MERGED | +2 | +1 |