Bug #52362 RQG not producing stack traces from MySQL upon declaring deadlock
Submitted: 25 Mar 2010 12:57 Modified: 13 Nov 2010 13:36
Reporter: John Embretsen Email Updates:
Status: No Feedback Impact on me:
None 
Category:Tools: Random Query Generator Severity:S3 (Non-critical)
Version: OS:Windows
Assigned to: Bernt Marius Johnsen CPU Architecture:Any
Tags: rqg_pb2

[25 Mar 2010 12:57] John Embretsen
Description:
This is a regression, as it used to work at some point in the not too distant past.

It seems that the RQG for some reason has stopped producing stack traces when declaring deadlock during a test run when the Deadlock reporter is enabled. 
It seems to be limited to the Windows OS, but this has not been verified 100% yet.

This is evidenced by recent runs of RQG MDL tests in Pushbuild, resulting in stalled queries (deadlocks). Logs show no stack traces. After the message

"6 stalled queries detected, declaring deadlock" (...)

a processlist dump is shown before gentest.pl exits, whereas it used to be followed by

"Server deadlock reported, initiating analysis..." and a cdb dump (on Windows). The latter is now missing.

Further investigation is needed to determine what change caused the regression (likely some time between March 09 and March 19).

How to repeat:
Run rqg_mdl_deadlock against mysql-trunk-runtime-exp until you see a deadlock(any branch or RQG test will probably do as long as it results in a deadlock declared by the Deadlock reporter). 

perl runall.pl \ 
 --grammar=conf/runtime/WL5004_sql.yy \ 
 --reporters=Deadlock,ErrorLog,Backtrace,Shutdown \ 
 --basedir=$CODE \ 
 --threads=10 \ 
 --queries=1M \ 
 --duration=1200 \ 
 --mysqld=--innodb \ 
 --mysqld=--innodb-lock-wait-timeout=50 \ 
 --mysqld=--lock-wait-timeout=31536000 \ 
 --mysqld=--log-output=file \ 
 --mysqld=--loose-skip-safemalloc \ 
 --mysqld=--loose-table-lock-wait-timeout=1
[13 Oct 2010 13:35] Bernt Marius Johnsen
Is it possible to check wether cdb is installed on the given machines?
[14 Nov 2010 0:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".