As doc states:

https://dev.mysql.com/doc/refman/5.7/en/innodb-page-compression.html

To disable page compression, set COMPRESSION=None using ALTER TABLE. Writes to the tablespace that occur after setting COMPRESSION=None no longer use page compression. To uncompress existing pages, you must rebuild the table using OPTIMIZE TABLE after setting COMPRESSION=None.

ALTER TABLE t1 COMPRESSION="None";
OPTIMZE TABLE t1;

With regular table there is no table rebuild:

mysql> alter table sbtest4 compression='none';
Query OK, 0 rows affected (0,01 sec)
Records: 0 Duplicates: 0 Warnings: 0

So the rebuild will be on optimize table run.

But if you want table rebuild each time when you run compression=’none’ you can add encryption=’N’ option

This post is about situation where comments in BUG reports are really valuable in term of learning and getting some internal view of going process.
Let’s take a look at -> http://bugs.mysql.com/bug.php?id=80183

The report is about how you can not drop general tablespace if it was a crash of MySQL instant while running alter statement.


mysql> create tablespace ts1 add datafile '/var/lib/mysql_tablespaces/ts1.ibd' engine=innodb;
Query OK, 0 rows affected (0,02 sec)

Kill -9 process while altering:


mysql> alter table sbtest1 tablespace ts1;
ERROR 2013 (HY000): Lost connection to MySQL server during query

CREATE TABLE `sbtest1` (
`id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`k` int(10) unsigned NOT NULL DEFAULT '0',
`c` char(120) NOT NULL …

How it feels when you realize that, something going to be scary in term of question – “Wait, why? how?”
Let’s see:

I am trying to allocate 9G buffer pool size with 64 instances:

innodb_buffer_pool_instances=64
innodb_buffer_pool_size=9G

From error log it said that:

InnoDB: Initializing buffer pool, total size = 16G, instances = 64, chunk size = 128M

“Wait, why? how?” N1

From select it is also allocated 16G:

mysql> select @@innodb_buffer_pool_size/1024/1024/1024;
+------------------------------------------+
| @@innodb_buffer_pool_size/1024/1024/1024 |
+------------------------------------------+
| 16.000000000000 |
+------------------------------------------+

Just continue to increase pool size:

If you give 17G:

[Note] InnoDB: Initializing buffer pool, …

If you are running ‘OPTIMIZE TABLE’ or altering ‘TABLESPACE’ just be careful not to kill MySQL service while running these commands
Here is a test case:
We have a table:

CREATE TABLE `sbtest1` (
`id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`k` int(10) unsigned NOT NULL DEFAULT '0',
`c` char(120) NOT NULL DEFAULT '',
`pad` char(60) NOT NULL DEFAULT '',
PRIMARY KEY (`id`),
KEY `k_1` (`k`)
) ENGINE=InnoDB AUTO_INCREMENT=2427738 DEFAULT CHARSET=latin1 COMPRESSION='lz4'

After a while we want to ‘OPTIMIZE’ this table:

mysql> optimize table sbtest1;
ERROR 2013 (HY000): Lost connection to MySQL server during query

Uups MySQL process was killed/crashed.

[root@ps-5 mysql]# ls -l dbtest/
total 1092792
-rw-r-----. 1 mysql mysql 65 фев 4 03:43 db.opt
-rw-r-----. 1 mysql mysql …

If you have not be careful with General Tablespaces in MySQL 5.7.10, you must be at least from now.
I have reported several bugs related to General Tablespaces for eg.:

* #77814 -> Altering table twice,with same tablespace will increase general tablespace size (verfied)

* #77556 -> Segmentation fault while creating general tablespace in disk full Linux (fixed/closed)

* #77830 -> “ALGORITHM=INPLACE is not supported” no error with tablespace alter (open)

* #80070 -> allocated_size and file_size differ if create general tablespace outside datadir (verified)

* …

Recently i have figured out that we can over-allocate InnoDB buffer pool size as I have already reported:

#79850

So as a result I can start MySQL with 300GB buffer pool size in my 16G laptop.

Config:

innodb_buffer_pool_size=300G
innodb_buffer_pool_instances=64

Started as:

2016-01-05 15:23:58 16211 [Note] InnoDB: Initializing buffer pool, size = 300.0G
2016-01-05 15:24:31 16211 [Note] InnoDB: Completed initialization of buffer pool

If you think that it is not harmful, let’s examine another report:

#79880

With MySQL 5.7.10 in my VM on CentOS 7:

I have:

[root@mysql-57 ~]# free -m
total used free shared buff/cache available
Mem: 3428 …

Well the new things are cool to test and learn. But it is not possible every time, to have smooth testing process. if you can’t get right software on your hand.

So there are great articles about this new plugin here you can follow all of them: Group Replication Topics

I must say that, plugin available for download as source code and as for Oracle Linux 6 x86_64 from -> labs.mysql.com

So if you want to try it with MySQL 5.7.8-rc2 on Ubuntu you should compile it with MySQL. But you will not be able to compile due to well known ‘boost’ dependency -> see BUG #78600

UPDATE to BUG 78600:

Pedro Gomes clarified the reason of …

What about inserting data to MySQL 5.7.7-rc-debug with GTID/binary log enabled. It is OK until you hit disk full condition while inserting data. It will crash on the way and likely you will lose MySQL. Error log indicates related problem:

2015-07-29T07:20:05.865708Z 2 [ERROR] InnoDB: posix_fallocate(): Failed to preallocate data for file ./sls2/sales.ibd, desired size 4194304 bytes. Operating system error number 28. Check that the disk is not full or a disk quota exceeded. Make sure the file system supports this function. Some operating system error numbers are described at http://dev.mysql.com/doc/refman/5.7/en/operating-system-error-codes.html
2015-07-29T07:20:05.895529Z 2 [Warning] InnoDB: 1048576 bytes should have been written. Only 557056 bytes written. Retrying again to write the remaining bytes.
2015-07-29T07:20:05.905149Z 2 [Warning] InnoDB: 491520 bytes should have been written. Only 106496 bytes written. Retrying again to write the remaining …

I am glad to share this post with community. It is really great story for a young guy to be among experts!

Today Lenka Kasparova announced results of: MySQL 5.7 Community Contributor Award Program 2015!

Awarded as:

Shahriyar Rzayev, the Azerbaijan MySQL User Group leader for reproducing bugs in MySQL 5.7

See my MySQL BUG database activity -> Reporter: Shahriyar Rzayev

Also there is a gift from MySQL Community team:

…

Dear community, maybe you have already tested, there are new added features in MySQL 5.7. See previous post -> About 5.7.7-rc BUGs

“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. Today’s topic is altering table to use this “General Tablespace”.

Our sample table structure:

mysql> show create table t1;

    | Table | Create Table 
    | t1    | CREATE TABLE `t1` (
      `id` int(11) DEFAULT NULL,
      `c1` varchar(85) DEFAULT NULL
    ) /*!50100 TABLESPACE innodb_file_per_table */ ENGINE=InnoDB DEFAULT CHARSET=utf8 

Note that there is a comment about tablespace of our table -> /*!50100 TABLESPACE …