[KUDU-2942] A rare flaky test for the aggregated live row count

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: test
• Assignee: Unassigned
• Reporter: LiFu He
• Created: 17/Sep/19 07:57
• Updated: 03/Jun/20 15:44

Description

A few days ago, Adar met a rare flaky test for the live row count in TSAN mode.

 

// code placeholder
/home/jenkins-slave/workspace/kudu-master/3/src/kudu/integration-tests/ts_tablet_manager-itest.cc:642
      Expected: live_row_count
      Which is: 327
To be equal to: table_info->GetMetrics()->live_row_count->value()
      Which is: 654

It seems the metric value is doubled. And his full test output is in the attachment.

 

I reviewed the previous patches and made some unusual guesses. I think one of them could explain the issue:

When one master just becomes the leader and there are two heartbeat messages from the same tserver that are processed in parallel at Line4239, then the metric value will be doubled because the old tablet stats can be accessed concurrently.

Thus, the question becomes how to generate two heartbeat messages from the same tserver at the same time? The possible answer is: First heartbeat message and Second heartbeat message

Please don't forget the above case is integrate test environment, not product.

 

 

 

Issue Links

• is related to: KUDU-2943 TsTabletManagerITest.TestTableStats flaky due to WAL/cmeta term disagreement
• relates to: KUDU-2944 TsTabletManagerITest.TestTableStats flaky due to data race