Description
The HmsSentryConfigurations/MasterStressTest seems to be a bit flaky if running via dist-test with --stress_cpu_threads=16. A snippet of the master-stress-test binary's output (DEBUG build) is below. It seems the configuration was
{{ HmsMode::ENABLE_METASTORE_INTEGRATION, SentryMode::DISABLED }}. Also, I'm attaching a full log.
I0426 04:05:42.127689 497 rpcz_store.cc:269] Call kudu.master.MasterService.AlterTable from 127.0.0.1:39526 (request call id 112) took 2593ms. Request Metrics: {"HiveMetastore.queue_time_us":1457223,"Hive Metastore.run_cpu_time_us":631,"HiveMetastore.run_wall_time_us":98149} F0426 04:05:42.132196 968 master-stress-test.cc:293] Check failed: _s.ok() Bad status: Remote error: failed to alter Hive MetaStore table: TException - service has thrown: MetaException
Attachments
Attachments
Issue Links
- duplicates
-
KUDU-2779 MasterStressTest is flaky when HMS is enabled
- Resolved