Describe the bug, including details regarding any error messages, version, and platform.
There are flakey mutex deadlock errors occurring in the Arrow Flight SQL ODBC tests after C++ 20 is enabled 30809c6, almost occurring in every test run. Mutex related changes have been made to enable C++ 20, so it could possibly be related.
I am only seeing this error in MSVC Windows, and I cannot reproduce it on MSYS2.
Acceptance Criteria:
- investigate, find root cause of the error and resolve it
Log:
The error has been seen in different tests as the appearance is random
[ RUN ] ConnectionAttributeTest/1.TestSQLSetConnectAttrTranslateLabDMOnly
[mutex.cc : 1426] RAW: Potential Mutex deadlock:
@ 00007FFF48880260
@ 00007FFF48878EB3
@ 00007FFF49A7C999
@ 00007FFF49BAC90A
@ 00007FFF4A05861C
@ 00007FFF4A01DF64
@ 00007FFF4A01D388
@ 00007FFF49CBD881
@ 00007FFF49CBE5B0
@ 00007FFF49CC2ACC
@ 00007FFF49CC18BC
@ 00007FFF49CC194D
@ 00007FFF49FF903E
@ 00007FFF49FF5332
@ 00007FFF49FF7F54
@ 00007FFF49FF756C
@ 00007FFF49FF78E1
@ 00007FFF4A0D16EA
@ 00007FFF4A7FDF8B
@ 00007FFF4A7F09C3
@ 00007FFF4A7F0519
@ 00007FFF4A7F1283
@ 00007FFF4A7F12E6
@ 00007FFF4A9C0B26
@ 00007FF84F09E8D7
@ 00007FF8500EC53C
[mutex.cc : 1438] RAW: Acquiring absl::Mutex 00000228D94F54B0 while holding 00000228D9C1D380; a cycle in the historical lock ordering graph has been observed
[mutex.cc : 1439] RAW: Cycle:
[mutex.cc : 1453] RAW: mutex@00000228D94F54B0 stack:
@ 00007FFF48880260
@ 00007FFF48878EB3
@ 00007FFF49A7C999
@ 00007FFF49BAC90A
@ 00007FFF4A05861C
@ 00007FFF4A40CE64
@ 00007FFF49FF2BA3
@ 00007FFF49FF28EC
@ 00007FFF49C9A141
@ 00007FFF49C9A33F
@ 00007FFF49C9A797
@ 00007FFF49C9CCF5
@ 00007FFF49C9DF9F
@ 00007FFF49E74EB5
@ 00007FFF49E73EA6
@ 00007FFF49CCB22D
@ 00007FFF49B63B81
@ 00007FFF49B348CA
@ 00007FFF49A48EBC
@ 00007FFF499D4B0A
@ 00007FF72955BF60
@ 00007FF72955CB04
@ 00007FFFA5966AED
@ 00007FFFA5966723
@ 00007FFFA592E76C
@ 00007FFFA592F474
@ 00007FFFA592FEF1
@ 00007FFFA59375C6
@ 00007FFFA5966BAD
@ 00007FFFA5966A33
@ 00007FFFA5930793
@ 00007FFFA59F2DE3
@ 00007FFFA59F2DB5
@ 00007FF729AE53F9
@ 00007FF729AE52A2
@ 00007FF729AE515E
@ 00007FF729AE548E
@ 00007FF84F09E8D7
@ 00007FF8500EC53C
[mutex.cc : 1453] RAW: mutex@00000228D9C1D380 stack:
@ 00007FFF48880260
@ 00007FFF48878EB3
@ 00007FFF4A9BCE74
@ 00007FFF4A2F2C4C
@ 00007FFF4A2F3146
@ 00007FFF4A2F412B
@ 00007FFF49DD8608
@ 00007FFF49BAD948
@ 00007FFF49BACE4B
@ 00007FFF4A05A3BC
@ 00007FFF4A40DAD8
@ 00007FFF4A426438
@ 00007FFF49C9D880
@ 00007FFF49C9DECD
@ 00007FFF499BA8F3
@ 00007FFF49B27584
@ 00007FFF49B240A4
@ 00007FFF4A685645
@ 00007FFF4A68B118
@ 00007FFF4A68A2E6
@ 00007FFF4A68B98B
@ 00007FFF499BA8F3
@ 00007FFF4A68B8E4
@ 00007FFF4A68AE24
@ 00007FFF4A683C97
@ 00007FFF4A7F0D0C
@ 00007FFF4A7F0519
@ 00007FFF4A7F1283
@ 00007FFF4A7F12E6
@ 00007FFF4A9C0B26
@ 00007FF84F09E8D7
@ 00007FF8500EC53C
[mutex.cc : 1461] RAW: dying due to potential deadlock
Parent issue: #30622
Component(s)
C++, FlightRPC
Describe the bug, including details regarding any error messages, version, and platform.
There are flakey mutex deadlock errors occurring in the Arrow Flight SQL ODBC tests after C++ 20 is enabled 30809c6, almost occurring in every test run. Mutex related changes have been made to enable C++ 20, so it could possibly be related.
I am only seeing this error in MSVC Windows, and I cannot reproduce it on MSYS2.
Acceptance Criteria:
Log:
The error has been seen in different tests as the appearance is random
Parent issue: #30622
Component(s)
C++, FlightRPC