This bug is very hard to reproduce. Established sporadic. When the testcase is reduced, if you execute the testcase at the CLI or via pquery, you end up with the stack from bug 1384656. However, using the reducer script itself, it is possible, even with a very small base testcase to produce this testcase. To do so;
#1 Change reducer125.sh - these variables (scroll about one page down from top of script);
#2 Excute reducer125.sh and observe output. It will start reproducing the bug. You can hack the script to stop at the required moments (simply search for output text in script) to for example allow time to attach debugger, or to write core file (add --core to MYEXTRA) etc.
[roel@localhost 883801]$ ./reducer125.sh
2014-11-05 01:59:12 [Init] Workdir: /dev/shm/1415113152
[...]
2014-11-05 01:59:28 [Init] Run mode: MODE3: mysqld error log
2014-11-05 01:59:28 [Init] Looking for this string: 'ZNK10MDL_ticket28has_pending_conflicting_lockEv' in mysqld error log output
[...]
2014-11-05 01:59:28 [] [Stage V] Verifying the bug/issue really exists (duration depends on initial input file size)
2014-11-05 01:59:28 [] [Stage V] [MULTI] Starting 10 verification subreducer threads to verify if the issue is sporadic (/dev/shm/1415113152/subreducer/)
2014-11-05 01:59:28 [] [Stage V] [MULTI] Ensuring any old subreducer processes are terminated
2014-11-05 01:59:30 [] [Stage V] [MULTI] Forking subreducer threads [PIDs]: #1 [18425] #2 [18682] #3 [18879] #4 [19024] #5 [19234] #6 [19389] #7 [19535] #8 [19699] #9 [19868] #10 [20029]
2014-11-05 01:59:30 [] [Stage V] [MULTI] Waiting for all forked verification subreducer threads to finish/terminate
2014-11-05 02:00:09 [] [Stage V] [MULTI] Finished/Terminated verification subreducer threads: #1 #2 #3 #4 #5 #6 #7 #8 #9 #10
2014-11-05 02:00:09 [] [Stage V] [MULTI] All verification subreducer threads have finished/terminated
2014-11-05 02:00:09 [*] [Stage V] [MULTI] Threads which reproduced the issue: #3 #4 #5 #7 #8 #10
2014-11-05 02:00:09 [*] [Stage V] [MULTI] Only 6 out of 10 threads reproduced the issue: this issue is sporadic
This bug is very hard to reproduce. Established sporadic. When the testcase is reduced, if you execute the testcase at the CLI or via pquery, you end up with the stack from bug 1384656. However, using the reducer script itself, it is possible, even with a very small base testcase to produce this testcase. To do so;
#1 Change reducer125.sh - these variables (scroll about one page down from top of script);
MYBASE= "/sda/Percona- Server- 5.6.21- rel69.0- 681.Linux. x86_64- debug" "/sda/883801/ 125/pquery_ thread- 0.sql_out_ out_out_ out_out_ out"
INPUTFILE=
Simply change path's to match your machine
#2 Excute reducer125.sh and observe output. It will start reproducing the bug. You can hack the script to stop at the required moments (simply search for output text in script) to for example allow time to attach debugger, or to write core file (add --core to MYEXTRA) etc.
[roel@localhost 883801]$ ./reducer125.sh ticket28has_ pending_ conflicting_ lockEv' in mysqld error log output 1415113152/ subreducer/ )
2014-11-05 01:59:12 [Init] Workdir: /dev/shm/1415113152
[...]
2014-11-05 01:59:28 [Init] Run mode: MODE3: mysqld error log
2014-11-05 01:59:28 [Init] Looking for this string: 'ZNK10MDL_
[...]
2014-11-05 01:59:28 [] [Stage V] Verifying the bug/issue really exists (duration depends on initial input file size)
2014-11-05 01:59:28 [] [Stage V] [MULTI] Starting 10 verification subreducer threads to verify if the issue is sporadic (/dev/shm/
2014-11-05 01:59:28 [] [Stage V] [MULTI] Ensuring any old subreducer processes are terminated
2014-11-05 01:59:30 [] [Stage V] [MULTI] Forking subreducer threads [PIDs]: #1 [18425] #2 [18682] #3 [18879] #4 [19024] #5 [19234] #6 [19389] #7 [19535] #8 [19699] #9 [19868] #10 [20029]
2014-11-05 01:59:30 [] [Stage V] [MULTI] Waiting for all forked verification subreducer threads to finish/terminate
2014-11-05 02:00:09 [] [Stage V] [MULTI] Finished/Terminated verification subreducer threads: #1 #2 #3 #4 #5 #6 #7 #8 #9 #10
2014-11-05 02:00:09 [] [Stage V] [MULTI] All verification subreducer threads have finished/terminated
2014-11-05 02:00:09 [*] [Stage V] [MULTI] Threads which reproduced the issue: #3 #4 #5 #7 #8 #10
2014-11-05 02:00:09 [*] [Stage V] [MULTI] Only 6 out of 10 threads reproduced the issue: this issue is sporadic