Testing MySQL 5.7.7-rc with memory and disc full conditions

With MySQL 5.7 there will be many improvements and changes. The most valuable ones are:

1. “General Tablespace” feature -> From now you can place tables to separate tablespaces. Prior, on our hands there were only “System TableSpace” and “InnoDB file per table” options.

Check for further reading: General Tablespaces Create statement: Create TableSpace

2. Resizing InnoDB Buffer Pool online -> for changing innodb_buffer_pool_size variable restart is required prior to MySQL 5.7, from now we can change buffer pool size without restart dynamically. Read Documentation -> innodb-buffer-pool-online-resize

To break this new things, we need to create some conditions where Memory or Disc will be full. Firstly let’s try to resize innodb buffer pool size without having sufficient memory.

Note: MySQL installation: debug enabled

Started MySQL as:

(gdb) run --defaults-file=/opt/mysql/my.cnf --basedir=/opt/mysql --datadir=/opt/mysql/datadir --plugin-dir=/opt/mysql/lib/plugin --user=mysql --log-error=/opt/mysql/datadir/centos7-1_vm.err --pid-file=/opt/mysql/datadir/mysqld-new.pid --socket=/opt/mysql/datadir/mysqld-new.sock --port=3306

Then try to resize:

mysql> SET @@GLOBAL.innodb_buffer_pool_size=9556544565345346;
Query OK, 0 rows affected, 2 warnings (0.00 sec)

After few attempts it will stuck with SIGABRT signal:

Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7fffb77f6700 (LWP 17145)]
0x00007ffff61675d7 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
56    return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);

(gdb) bt
#0  0x00007ffff61675d7 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#1  0x00007ffff6168cc8 in __GI_abort () at abort.c:90
#2  0x0000000001adbe0f in ut_dbg_assertion_failed (expr=0x0, file=0x21956d8 "/root/mysql-5.7.7-rc/storage/innobase/ut/ut0ut.cc", line=906)
at /root/mysql-5.7.7-rc/storage/innobase/ut/ut0dbg.cc:67
#3  0x0000000001ae123e in ib::fatal::~fatal (this=0x7fffb77f49e0, __in_chrg=<optimized out>) at /root/mysql-5.7.7-rc/storage/innobase/ut/ut0ut.cc:906
#4  0x00000000018f18e8 in ut_allocator<unsigned char>::allocate (this=0x7fffb77f4ba0, n_elements=14238670808, hint=0x0, 
    file=0x21cbb10 "/root/mysql-5.7.7-rc/storage/innobase/ha/hash0hash.cc", set_to_zero=false, throw_on_error=false)
    at /root/mysql-5.7.7-rc/storage/innobase/include/ut0new.h:328
#5  0x0000000001bed775 in hash0_create (n=1756528640) at /root/mysql-5.7.7-rc/storage/innobase/ha/hash0hash.cc:284
#6  0x000000000194769d in lock_sys_resize (n_cells=1756528640) at /root/mysql-5.7.7-rc/storage/innobase/lock/lock0lock.cc:478
#7  0x0000000001b3dce1 in buf_pool_resize () at /root/mysql-5.7.7-rc/storage/innobase/buf/buf0buf.cc:2669
#8  0x0000000001b3e0f5 in buf_resize_thread (arg=0x0) at /root/mysql-5.7.7-rc/storage/innobase/buf/buf0buf.cc:2749
#9  0x00007ffff7bc6df5 in start_thread (arg=0x7fffb77f6700) at pthread_create.c:308
#10 0x00007ffff62281ad in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113

See Related verified BUG report -> #77564

How about creating “General Tablespace” if there is no free disk space? Again started with:

[root@centos7_vm ~]# gdb /opt/mysql/bin/mysqld
(gdb) run --defaults-file=/opt/mysql/my.cnf --basedir=/opt/mysql --datadir=/opt/mysql/datadir --plugin-dir=/opt/mysql/lib/plugin --user=mysql --log-error=/opt/mysql/datadir/centos7_vm.err --pid-file=/opt/mysql/datadir/mysqld-new.pid --socket=/opt/mysql/datadir/mysqld-new.sock --port=3306

Try to run create tablespace:

mysql> CREATE TABLESPACE `ts2` ADD DATAFILE 'ts2.ibd' Engine=InnoDB;

In error log there is some mention about full disk:

2015-06-30T05:34:41.379790Z 2 [ERROR] InnoDB: Encountered a problem with file ./ts2.ibd
2015-06-30T05:34:41.379865Z 2 [ERROR] InnoDB: Disk is full. Try to clean the disk to free space.
2015-06-30T05:34:41.379910Z 2 [ERROR] InnoDB: Operating system error number 28 in a file operation.
2015-06-30T05:34:41.379937Z 2 [ERROR] InnoDB: Error number 28 means 'No space left on device'.
2015-06-30T05:34:41.379952Z 2 [Note] InnoDB: Some operating system error numbers are described at http://dev.mysql.com/doc/refman/5.7/en/operating-system-error-codes.html
2015-06-30T05:34:41.393247Z 2 [ERROR] InnoDB: Cannot create file './ts2.ibd'

From GDB we will see SIGSEGV signal:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffddf7e700 (LWP 2787)]
__strnlen_sse2 () at ../sysdeps/x86_64/strnlen.S:34
34      pcmpeqb (%rdi), %xmm2

    (gdb) bt
#0  __strnlen_sse2 () at ../sysdeps/x86_64/strnlen.S:34
#1  0x0000000001dc18f2 in process_str_arg (cs=0x2ba3ea0 <my_charset_utf8_general_ci>, to=0x7fffddf7bfdb "02", end=0x7fffddf7c1cf "", width=192, 
    par=0x1 <Address 0x1 out of bounds>, print_type=0) at /root/mysql-5.7.7-rc/strings/my_vsnprintf.c:194
#2  0x0000000001dc32e6 in my_vsnprintf_ex (cs=0x2ba3ea0 <my_charset_utf8_general_ci>, to=0x7fffddf7bfdb "02", n=512, fmt=0x2d32a12 "s' is full", 
    ap=0x7fffddf7c1d0) at /root/mysql-5.7.7-rc/strings/my_vsnprintf.c:607
#3  0x000000000189d0e9 in my_error (nr=135, MyFlags=0) at /root/mysql-5.7.7-rc/mysys/my_error.c:206
    #4  0x0000000001622428 in mysql_alter_tablespace (thd=0x7fffcc000e20, ts_info=0x7fffcc022080) at /root/mysql-5.7.7-rc/sql/sql_tablespace.cc:191
#5  0x000000000158e104 in mysql_execute_command (thd=0x7fffcc000e20) at /root/mysql-5.7.7-rc/sql/sql_parse.cc:4432
#6  0x000000000158ff62 in mysql_parse (thd=0x7fffcc000e20, parser_state=0x7fffddf7d5a0) at /root/mysql-5.7.7-rc/sql/sql_parse.cc:5159
#7  0x0000000001585cf1 in dispatch_command (command=COM_QUERY, thd=0x7fffcc000e20, 
packet=0x7fffcc006f71 "CREATE TABLESPACE `ts2` ADD DATAFILE 'ts2.ibd' Engine=InnoDB", packet_length=60) at /root/mysql-5.7.7-rc/sql/sql_parse.cc:1249
#8  0x0000000001584b21 in do_command (thd=0x7fffcc000e20) at /root/mysql-5.7.7-rc/sql/sql_parse.cc:835
#9  0x00000000016adcff in handle_connection (arg=0x3c92060) at /root/mysql-5.7.7-rc/sql/conn_handler/connection_handler_per_thread.cc:298
#10 0x0000000001cd9905 in pfs_spawn_thread (arg=0x3b4dcd0) at /root/mysql-5.7.7-rc/storage/perfschema/pfs.cc:2147
#11 0x00007ffff7bc6df5 in start_thread (arg=0x7fffddf7e700) at pthread_create.c:308
#12 0x00007ffff62281ad in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113

Related BUG report -> #77556

If you have you own test cases, please share it.

Author: Shahriyar Rzayev

Azerbaijan MySQL User Group and Python user group leader. QA Engineer, bug hunter by nature and true Pythonista

2 thoughts on “Testing MySQL 5.7.7-rc with memory and disc full conditions”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.