ALTER TABLE ... ENGINE = TOKUDB crashes on Ubuntu 12.04

[Sveta Smirnova]

ALTER TABLE ... ENGINE = TOKUDB crashes on Ubuntu 12.04

How to repeat:

0. Fresh Ubuntu Precise64 install:

$ uname -a
Linux precise64 3.2.0-23-generic #36-Ubuntu SMP Tue Apr 10 20:39:51 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

1. Download latest 5.6 version: Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100

2. Install jemalloc, libaio, libnuma, disable transparent pages:

vagrant@precise64:~/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test$ sudo su
root@precise64:/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test# echo never > /sys/kernel/mm/transparent_hugepage/enabled
root@precise64:/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test# echo never > /sys/kernel/mm/transparent_hugepage/defrag
root@precise64:/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test# exit

3. Start MTR:

LD_PRELOAD=/usr/lib/libjemalloc.so ./mtr --start --suite=tokudb --mysqld=--log-bin=master auto_increment &

4. Enable TokuDB:

vagrant@precise64:~/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test$ sudo ../bin/ps_tokudb_admin --enable -uroot -h127.0.0.1 -P13001
Checking if Percona Server is running with jemalloc enabled...
INFO: Percona Server is running with jemalloc enabled.

Checking transparent huge pages status on the system...
INFO: Transparent huge pages are enabled (should be disabled).

Checking if thp-setting=never option is already set in config file...
INFO: Option thp-setting=never is set in the config file.

Checking TokuDB engine plugin status...
ERROR: TokuDB engine plugin is partially installed. Please cleanup manually.

First strange thing: it complains about transparent pages, but they are disabled

5. Verify TokuDB engine enabled:

mysql> show engines;
+--------------------+---------+----------------------------------------------------------------------------+--------------+------+------------+
| Engine | Support | Comment | Transactions | XA | Savepoints |
+--------------------+---------+----------------------------------------------------------------------------+--------------+------+------------+
| InnoDB | YES | Percona-XtraDB, Supports transactions, row-level locking, and foreign keys | YES | YES | YES |
| MRG_MYISAM | YES | Collection of identical MyISAM tables | NO | NO | NO |
| MyISAM | DEFAULT | MyISAM storage engine | NO | NO | NO |
| BLACKHOLE | YES | /dev/null storage engine (anything you write to it disappears) | NO | NO | NO |
| CSV | YES | CSV storage engine | NO | NO | NO |
| TokuDB | YES | Percona TokuDB Storage Engine with Fractal Tree(tm) Technology | YES | YES | YES |
...

6. create database sbtest;

7. Create a table (non-TokuDB):

sysbench --test=oltp --oltp-table-size=1000000 --mysql-user=root --mysql-host=127.0.0.1 --mysql-port=13001 prepare

8. Alter sbtest table:

mysql> alter table sbtest engine=tokudb;
ERROR 2013 (HY000): Lost connection to MySQL server during query
mysql> worker[1] Server [mysqld.1 - pid: 8200, winpid: 8200, exit: 256] died

Only 0 of 1 completed.
mysql-test-run: *** ERROR: Not all tests completed

Relevant part of the error log file:

Version: '5.6.30-76.3-log' socket: '/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/mysql-test/var/tmp/mysqld.1.sock' port: 13001 Percona Server (GPL), Release 76.3, Revision 3850db5
21:30:04 UTC - mysqld got signal 11 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.
Please help us make Percona Server better by reporting any
bugs at http://bugs.percona.com/

key_buffer_size=1048576
read_buffer_size=131072
max_used_connections=1
max_threads=153
thread_count=1
connection_count=1
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 61881 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x7f1914b5f000
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 7f19218bae10 thread_stack 0x40000
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(my_print_stacktrace+0x2e)[0x8e087e]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(handle_fatal_signal+0x491)[0x6a7881]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xfcb0)[0x7f19212d7cb0]
/usr/lib/libjemalloc.so(free+0x6b)[0x7f19214e8e1b]
/lib/x86_64-linux-gnu/libpthread.so.0(pthread_create+0x279)[0x7f19212d03b9]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/lib/mysql/plugin/ha_tokudb.so(_Z20toku_ft_loader_closeP11ft_loader_sPFvP9__toku_dbiiP10__toku_dbtS4_PvES5_PFiS5_fES5_+0x3a2)[0x7f191957d982]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/lib/mysql/plugin/ha_tokudb.so(_Z17toku_loader_closeP13__toku_loader+0x157)[0x7f191957e8f7]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/lib/mysql/plugin/ha_tokudb.so(_ZN9ha_tokudb15end_bulk_insertEb+0x4be)[0x7f19194abd1e]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_Z17mysql_alter_tableP3THDPcS1_P24st_ha_create_informationP10TABLE_LISTP10Alter_infojP8st_orderb+0x2c4e)[0x769c6e]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_ZN19Sql_cmd_alter_table7executeEP3THD+0x43c)[0x83445c]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_Z21mysql_execute_commandP3THD+0xd9a)[0x71a95a]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_Z11mysql_parseP3THDPcjP12Parser_state+0x598)[0x71ef58]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcj+0xef8)[0x71fee8]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(_Z24do_handle_one_connectionP3THD+0x15f)[0x6f392f]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(handle_one_connection+0x47)[0x6f3a17]
/home/vagrant/Percona-Server-5.6.30-rel76.3-Linux.x86_64.ssl100/bin/mysqld(pfs_spawn_thread+0x143)[0xb3be23]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e9a)[0x7f19212cfe9a]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d)[0x7f191f5604bd]

Trying to get some variables.
Some pointers may be invalid and cause the dump to abort.
Query (7f18fb01b010): alter table sbtest engine=tokudb
Connection ID (thread ID): 5
Status: NOT_KILLED

[Sveta Smirnova (svetasmirnova)]

Bug is not repeatable on Ubuntu 14.04

[Sveta Smirnova (svetasmirnova)]

Reason for the failure is outdated system jemalloc. 5.6.30 and 5.7.12 include copy of jemalloc library in *tar.gz packages LD_PRELOAD should point to this library.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-992