[MDEV-30242] MTR fails to report stack traces of all threads by default Created: 2022-12-16  Updated: 2022-12-16  Resolved: 2022-12-16

Status: Closed
Project: MariaDB Server
Component/s: Tests
Affects Version/s: 10.3.36, 10.4.26, 10.5.17, 10.6.9, 10.7.5, 10.8.4, 10.9.2, 10.10.1, 10.3.37, 10.4.27, 10.5.18, 10.6.10, 10.7.6, 10.8.5, 10.9.3, 10.6.11, 10.7.7, 10.8.6, 10.9.4, 10.10.2
Fix Version/s: 10.11.2, 10.3.38, 10.4.28, 10.5.19, 10.6.12, 10.7.8, 10.8.7, 10.9.5, 10.10.3

Type: Bug Priority: Major
Reporter: Marko Mäkelä Assignee: Marko Mäkelä
Resolution: Fixed Votes: 0
Labels: regression

Issue Links:
Problem/Incident
is caused by MDEV-28931 MTR prints detailed stack trace uncon... Closed
Relates
relates to MDEV-30084 Shutdown hangs in some tests Closed

 Description   

MDEV-28931 implemented a change that made it impossible by default to diagnose many race conditions or hangs based on output that is available from CI systems. MDEV-30084 is an example of that: A test would fail rarely, and we had no idea about the root cause until the insufficient GDB command bt (short for backtrace) was replaced with the more appropriate thread apply all backtrace, so that the stack traces of all active threads were displayed.

It may be more convenient for interactive users of the test script if the crash reports are as short as possible, because interactive users usually can access core dumps or rerun failed tests afterwards. That is not possible on CI systems.

It is easier for individual users to adapt the way how they invoke the regression test suite than for all CI systems out there to be adjusted, to "opt in" for more detailed reporting of crashes.

Let us re-enable the output of stack traces all threads by default. Users who do not like verbose output can set the environment variable MTR_PRINT_CORE=small or invoke ./mtr --skip-core-file or similar.


Generated at Thu Feb 08 10:14:46 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.