Bug #48470 ndb_restore crash specific to one platform
Submitted: 2 Nov 2009 13:12
Reporter: Joerg Bruehe Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:mysql-5.1-telco-6.3 OS:Mac OS X (10.5 (x86: 32 bit))
Assigned to: CPU Architecture:Any
Tags: cluster-6.3.27a
Triage: Triaged: D2 (Serious) / R6 (Needs Assessment) / E6 (Needs Assessment)

[2 Nov 2009 13:12] Joerg Bruehe
Description:
The following crash is new in the build of 6.3.27a,
at least I do not find in in either the bugs DB or in previous isolated test failures:

=====
parts.ndb_dd_backuprestore               [ retry-fail ]
        Test ended at YYYY-MM-DD HH:MM:SS

CURRENT_TEST: parts.ndb_dd_backuprestore
sh: line 1: PID DIED   /Users/mysqldev/tmp-200910271837-6.3.27a-6980/osx10.5-x86/test/bin//ndb_restore --no-defaults --ndb-connectstring="localhost:13000" -p 8 -b 2 -n 1 -m -r --print --print_meta /Users/mysqldev/tmp-200910271837-6.3.27a-6980/osx10.5-x86/test/mysql-test/var/mysql_cluster.1//BACKUP/BACKUP-2 >> /Users/mysqldev/tmp-200910271837-6.3.27a-6980/osx10.5-x86/test/mysql-test/var/log/ndb_testrun.log
mysqltest: In included file "./include/ndb_restore_master.inc": At line 6: command "$NDB_TOOLS_DIR/ndb_restore
 --no-defaults --ndb-connectstring="$NDB_CONNECTSTRING" -p 8 -b $the_backup_id -n 1 -m -r --print --print_meta
 $NDB_BACKUP_DIR/BACKUP/BACKUP-$the_backup_id >> $NDB_TOOLS_OUTPUT" failed

Output from before failure:
exec of '/PATH/bin//ndb_restore --no-defaults --ndb-connectstring="localhost:13000" -p 8 -b 2 -n 1 -m -r --pri
nt --print_meta /Users/mysqldev/tmp-200910271837-6.3.27a-6980/osx10.5-x86/test/mysql-test/var/mysql_cluster.1/
/BACKUP/BACKUP-2 >> /Users/mysqldev/tmp-200910271837-6.3.27a-6980/osx10.5-x86/test/mysql-test/var/log/ndb_test
run.log' failed, error: 35584, status: 139, errno: 2

The result from queries just before the failure was:
< snip >
SELECT COUNT(*) FROM test.t4;
COUNT(*)
100
SELECT LENGTH(data) FROM test.t4 WHERE c1 = 1;
LENGTH(data)
1024
SELECT LENGTH(data) FROM test.t4 WHERE c1 = 2;
LENGTH(data)
16384
DROP TABLE test.t1;
DROP TABLE test.t2;
DROP TABLE test.t3;
DROP TABLE test.t4;
ALTER TABLESPACE table_space1
DROP DATAFILE './table_space1/datafile.dat'
ENGINE = NDB;
DROP TABLESPACE table_space1
ENGINE = NDB;
DROP LOGFILE GROUP log_group1
ENGINE =NDB;

More results from queries before failure can be found in /PATH/mysql-test/var/log/ndb_dd_backuprestore.log
=====

It happens with all three configurations ("gpl", "com", "com-pro") on just that one platform.
I read status 139 to mean SIGSEGV (signal 11) + core dump (128).

How to repeat:
Happened while running the full test suite.
[3 Nov 2009 14:18] Jørgen Austvik
Could you please provide a stack trace for the core dump of the crashed process?