MySQL Bugs: #16550: Cluster Replication: rpl_create_database.test cause core on slave MySQLD

Bug #16550	Cluster Replication: rpl_create_database.test cause core on slave MySQLD
Submitted:	16 Jan 2006 22:09	Modified:	17 Jan 2006 8:39
Reporter:	Jonathan Miller	Email Updates:
Status:	Can't repeat	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S1 (Critical)
Version:	5.1.6-alpha	OS:	Linux (Linux)
Assigned to:	Tomas Ulin	CPU Architecture:	Any

Description:
rpl_create_database            [ fail ]

Errors are (from /home/ndbdev/jmiller/clones/mysql-5.1-new/mysql-test/var/log/mysqltest-time) :
mysqltest: At line 46: failed in select master_pos_wait('master-bin.000001', 1198): 2013: Lost connection to MySQL server during query
#0  0x00b4e402 in __kernel_vsyscall ()
#1  0x00a2855f in pthread_kill () from /lib/libpthread.so.0
#2  0x082dbbdb in write_core (sig=454) at stacktrace.c:220
#3  0x081c2e22 in handle_segfault (sig=11) at mysqld.cc:2181
#4  <signal handler called>
#5  0x08502a6a in NdbTransaction::execute (this=0x0, aTypeOfExec=Commit,
    abortOption=AbortOnError, forceSend=1) at NdbTransaction.cpp:267
#6  0x083304dc in ndbcluster_commit (thd=0x9f2f9a0, all=false) at ha_ndbcluster.cc:279
#7  0x082916b2 in ha_commit_one_phase (thd=0x9f2f9a0, all=false) at handler.cc:741
#8  0x082928e5 in ha_commit_trans (thd=0x9f2f9a0, all=false) at handler.cc:711
#9  0x08292b06 in ha_autocommit_or_rollback (thd=0x9f2f9a0, error=0) at handler.cc:850
#10 0x08259a62 in Rows_log_event::exec_event (this=0x9f55b28, rli=0xb7a20438)
    at log_event.cc:5477
#11 0x082cf515 in handle_slave_sql (arg=0xb7a1f648) at slave.cc:3047
#12 0x00a25b80 in start_thread () from /lib/libpthread.so.0
#13 0x002e89ce in clone () from /lib/libc.so.6

How to repeat:
./mysql-test-run  --mysqld=--binlog-format=row --do-test=rpl --with-ndbcluster-all --mysqld=--default-storage-engine=ndb --force --with-ndbcluster-slave

can't repeat this by running the test standalone

also cannot repeat when run the whole rpl test suite

I've run all rpl tests with ndb storage engine

Failed 60/122 tests, 50.81% were successful.

many seem to be just order problem in result set

in debug mode I get 2 slave cores

- one is a known bug with "create table like", which only will  be detected in debug mode
- the other is a safe_mutex_lock in process alarm which seems unrelated to replication

#10 0x40239e60 in abort () from /lib/libc.so.6
#11 0x08674913 in safe_mutex_lock (mp=0x88e7aa0, file=0x87b9700 "thr_alarm.c", line=298) at thr_mutex.c:104
#12 0x0867182a in process_alarm (sig=14) at thr_alarm.c:298
#13 0x40040d69 in __pthread_clock_settime () from /lib/libpthread.so.0
#14 <signal handler called>
#15 0x40238d8f in sigsuspend () from /lib/libc.so.6
#16 0x4003e8fa in sigwait () from /lib/libpthread.so.0

marking as can't repeat

also could not make out anything from the core other that that a commit on a transaction is attempted on the slave, although the ndb handler has no recollaction of an ongoing transaction