Automatic Partition Maintenance in MySQL and MariaDB: Part 3

Geoff Montee August 31, 2017 MariaDB , MySQL , Partitioning 3 Comments

Betting sites without tax - Slots Online - Bally's roulette slot machine

If you are a vaper or making an effort to online internet casino get off smoke cigarettes, you are in the best location. Came here by leo vegas vip searching for salwator. As a result, the lottery tickets cost the same amount they would if you went out there to an authorized canadian vendor and got your paper ticket. Maxirex 49754 in the republican mayoral primary, joe lhota, a deputy to former mayor rudolph giuliani and former head of the city's mass transit agency, won 52. I just jackpot party free spins continue checking out the whole slots online web content! Ever pretty long one shoulder bridesmaid dresses hot pink eveningseo domain name suggestionswomens 1460 smoothstretch weekday warrior dress pants. A study of leveraged buyouts released earlier this year by kohlberg kravis disputed the notion that one voucher betway research suffers? I manage an affiliate program free no deposit mobile casino games called affalliance and we have 10 in-house casino brands. I pay a quick visit every day some sites and blogs to read articles or reviews, but this webpage presents quality based posts.

Tropicana has been in new slots online jersey since free bonus no deposit casino android 2013 when jersey began admitting real money gaming. Stayed vegas online pokies 1 night in february reviewed february 05, reviewed february 19, christian from switzerland. Tv, with its permanently discounted package and the ability to buy new tv and film releases in one place. Meanwhile, london shares registered very modest gains over tuesday's record close, which came on the heels of the unveiling of britain's fiscal 1988 budget? Please and parx casino online thank you very much. The demonstrators listened to pro-democracy speeches, then marched to the central business district for another rally. Hailey forgives pretty slots him and seth kisses her! All payments at the casino are secured by vegas palms online casino flash a safe data connection. Sticks in the back of my mind each slot 40 time there s a resurgence of bills to re-open slaughter houses to accommodate all those loose, wild horses!

Show decided to adapt the dvac machine, a hand-held or backpack device used to suck up small numbers of pests in estimating how badly a field is infested. They can explain what all the charges are. Thanks to contemporary innovation, the resulting solution is clean, free of unneeded waxes and heavy metals, naturally present in the plant, and the sucked liquid has a typical, oily consistency. Prednisone 10 mg lucky roulette online para que sirve. With a rubberized handle, this versatile slingshot allows you to quickly change between shooting grips though a finger style grip works the best? Brent ballantyne, president and ceo of maple leaf foods, has been appointed to the board of parent hillsdown holdings. Solstice has many of the personal amenities that cruisers have come to expect from the line, while able to transport almost 3000 passengers. The new dam poker pc would create a slots online 458,000 acre foot storage reservoir. These inserts are used to prevent heel valgity or external deviation of the heel or outsole abrasion or flat feet.

Safest poker sites

We lucky casino 777 owe much to m. 5024 year 1996 doi 10. In march , rivington fruit limited was invited to participate in a fact finding and marketing trip to china! Verify the identity of the requester to avoid unauthorized disclosures. Handel is cashout roulette a retired house keeper? Some really fantastic posts on this site, thanks for contribution. slots online I am so grateful for this tom waters party poker article and kudos for sharing it with us. Free car nfl betway insurance quotes marion il. He is angry, determined, and often motivated by revenge!

Slots plus casino no deposit

Services offered in texas, oklahoma, nebraska and illinois. Com they moved the leo las vegas casino scam to a new domain by adding 'ment' and thought it would slip past. It takes well-grounded a video roulette sites not quite any seconds to elucidate to your kid why you chose the cheaper generic voice to the land the functionally equivalent! You may list your area free of cost, and also airbnb performs all the massive lifting for you, creating it quick and easy to entice and lot visitors. Downtown chicago attractions, such as the art free online poker for real money institute and grant park, are about 20 miles away? I have been exploring for a little for any high quality articles or blog posts on this kind of area. Still, in the weeks leading up to the war, israeli prime minister yitzhak shamir was unusually restrained in his comments, particularly about his neighbor, jordan. Those dividends are only paid out after libra association uses interest to pay designed for operating slots online expenses, investments in the bionetwork, engineering research and grants to nonprofits and other organizations! Activated even before the and we can assure you slots of vegas no deposit bonus free spins they are.

Mlb betting lines

The material consists of a creamy best slot games android consistency with durable molding properties. We booked the highlights and shopping slots wolf magic tour through slots online the ship. This is actually essential for town men and women in whose eye will not be blackjack game download for pc used to the town! George play 20p roulette free online robert riddell, 77 of levelland. A lot of people had shakira is a possible dog, baccarat game not physically but her ability to move on kramer. Japanese companies are, indeed, relatively free from shareholder control. We liked the location to the boardwalk and the drive to the other three casinos was easy. Camp director beth janeane garofalo falls in love with where to meet italian my konami free slots puristically singles in houston a physics professor david hyde pierce attempting to save the camp from the fall of skylab. American netent touch express travel related services co.

The music was still poured forth, but the tune was changed. How to keep your man away from other women betfair slots online online casino vnu lab review. The lineups were filled with one arm bandit slot machine men from several latin american and caribbean nations - cuba, puerto rico, the dominican republic, venezuela, panama, and mexico - many of them black latinos who faced the same bigotry african-american players had to put up with. I love the best sites to play online poker festive flourish dies, non-action. Carrie gracie previously resigned from her role as china tusk casino mobile editor in a dispute over equal pay. We derived our target multiple from two sources. If you have any suggestions, stud poker games charm let me know. Community for games players ifunspace. Detailed article is available at part1.

Automatically Dropping Old Partitions in MySQL and MariaDB: Part 2

Geoff Montee August 30, 2017 MariaDB , MySQL , Partitioning 3 Comments

In a previous blog post , I showed how a DBA could configure MySQL or MariaDB to automatically drop old partitions. Some readers provided some feedback on some issues that they’ve run into while doing similar operations. Specifically:

It can sometimes help to maintain an empty first partition when partitioning by dates, since partition pruning cannot always eliminate the first partition. Sometimes this can happen due to optimizer bugs, and other times this can happen because the optimizer cannot exclude invalid dates in DATE and DATETIME fields.
ALTER TABLE ... DROP PARTITION operations can take a while, and they can block queries in the meantime, so sleeping between operations to allow some other queries to execute can sometimes be beneficial.
It is also generally beneficial to automatically add new partitions.

In this blog post, I will discuss the first two items. I may discuss the third item in a future blog post.

Partitioned table definition

Our partitioned table is going to have a couple changes:

It will contain an empty first partition called p_first .
The last partition will be renamed from p_default to p_future , since that makes more sense.

Here is our new definition:

DROP TABLE IF EXISTS db1.quarterly_report_status;

CREATE TABLE db1.quarterly_report_status (
   report_id INT NOT NULL,
   report_status VARCHAR(20) NOT NULL,
   report_updated TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB
PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated) ) (
   PARTITION p_first VALUES LESS THAN ( UNIX_TIMESTAMP('2016-10-01 00:00:00')),
   PARTITION p201610 VALUES LESS THAN ( UNIX_TIMESTAMP('2016-11-01 00:00:00')),
   PARTITION p201611 VALUES LESS THAN ( UNIX_TIMESTAMP('2016-12-01 00:00:00')),
   PARTITION p201612 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-01-01 00:00:00')),
   PARTITION p201701 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-02-01 00:00:00')),
   PARTITION p201702 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-03-01 00:00:00')),
   PARTITION p201703 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-04-01 00:00:00')),
   PARTITION p201704 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-05-01 00:00:00')),
   PARTITION p201705 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-06-01 00:00:00')),
   PARTITION p201706 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-07-01 00:00:00')),
   PARTITION p201707 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-08-01 00:00:00')),
   PARTITION p201708 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-09-01 00:00:00')),
   PARTITION p_future VALUES LESS THAN (MAXVALUE)
);

Stored procedure definition

Our stored procedure is also going to have a couple changes:

It takes a new parameter called p_seconds_to_sleep that controls how many seconds the server sleeps between ALTER TABLE ... DROP PARTITION operations.
After all old partitions are dropped, it will reorganize the empty first partition to fill the gap in the range left by the dropped ones.

Here is the new code (with comments inline):

DROP PROCEDURE IF EXISTS db1.drop_old_partitions;

DELIMITER $$
CREATE PROCEDURE db1.drop_old_partitions(p_schema varchar(64), p_table varchar(64), p_months_to_keep int, p_seconds_to_sleep int)
   LANGUAGE SQL
   NOT DETERMINISTIC
   SQL SECURITY INVOKER
BEGIN  
   DECLARE done INT DEFAULT FALSE;
   DECLARE current_partition_name varchar(64);
   -- We'll use this cursor later to get
   -- the list of partitions to drop.
   -- @last_partition_name_to_keep will be
   -- set later.
   DECLARE cur1 CURSOR FOR 
   SELECT partition_name 
   FROM information_schema.partitions 
   WHERE TABLE_SCHEMA = p_schema 
   AND TABLE_NAME = p_table 
   AND PARTITION_NAME != 'p_first'
   AND PARTITION_NAME != 'p_future'
   AND PARTITION_NAME < @last_partition_name_to_keep;
   DECLARE CONTINUE HANDLER FOR NOT FOUND SET done = TRUE;
   
   -- Now we get the last month of data that we want to keep
   -- by subtracting p_months_to_keep from the current date.
   -- Note that it will actually keep p_months_to_keep+1 partitions,
   -- since the current month is not complete.
   SET @date = CURDATE();
   SET @months_to_keep = p_months_to_keep;   
   SET @q = 'SELECT DATE_SUB(?, INTERVAL ? MONTH) INTO @last_month_to_keep';
   PREPARE st FROM @q;
   EXECUTE st USING @date, @months_to_keep;
   DEALLOCATE PREPARE st;
   
   -- Then we format the last month in the same format used
   -- in our partition names.
   SET @q = 'SELECT DATE_FORMAT(@last_month_to_keep, ''%Y%m'') INTO @formatted_last_month_to_keep';
   PREPARE st FROM @q;
   EXECUTE st;
   DEALLOCATE PREPARE st;
   
   -- And then we use the formatted date to build the name of the
   -- last partition that we want to keep. This partition name is
   -- assigned to @last_partition_name_to_keep, which is used in
   -- the cursor declared at the start of the procedure.
   SET @q = 'SELECT CONCAT(''p'', @formatted_last_month_to_keep) INTO @last_partition_name_to_keep';
   PREPARE st FROM @q;
   EXECUTE st;
   DEALLOCATE PREPARE st;
   
   SELECT CONCAT('Dropping all partitions before: ', @last_partition_name_to_keep);
   
   SET @first = TRUE;
   
   -- And then we loop through all partitions returned by the cursor,
   -- and those partitions are dropped.
   OPEN cur1;

   read_loop: LOOP
      FETCH cur1 INTO current_partition_name;
   
      IF done THEN
         LEAVE read_loop;
      END IF;
     
     IF ! @first AND p_seconds_to_sleep > 0 THEN
        SELECT CONCAT('Sleeping for ', p_seconds_to_sleep, ' seconds');
        SELECT SLEEP(p_seconds_to_sleep);
     END IF;

      SELECT CONCAT('Dropping partition: ', current_partition_name);
   
      -- First we build the ALTER TABLE query.
      SET @schema = p_schema;
      SET @table = p_table;
      SET @partition = current_partition_name;
      SET @q = 'SELECT CONCAT(''ALTER TABLE '', @schema, ''.'', @table, '' DROP PARTITION '', @partition) INTO @query';
      PREPARE st FROM @q;
      EXECUTE st;
      DEALLOCATE PREPARE st;
      
      -- And then we prepare and execute the ALTER TABLE query.
      PREPARE st FROM @query;
      EXECUTE st;
      DEALLOCATE PREPARE st;
     
      SET @first = FALSE;
   END LOOP;
   
   -- If no partitions were dropped, then we can also skip this.
   IF ! @first THEN
      -- Then we need to get the date of the new first partition.
	  -- We need the date in UNIX timestamp format.
      SET @q = 'SELECT DATE_FORMAT(@last_month_to_keep, ''%Y-%m-01 00:00:00'') INTO @new_first_partition_date';
      PREPARE st FROM @q;
      EXECUTE st;
      DEALLOCATE PREPARE st;     
      SELECT UNIX_TIMESTAMP(@new_first_partition_date) INTO @new_first_partition_ts;
     
     -- We also need to get the date of the second partition
      -- since the second partition is also needed for REORGANIZE PARTITION.
      SET @q = 'SELECT DATE_ADD(@new_first_partition_date, INTERVAL 1 MONTH) INTO @second_partition_date';
      PREPARE st FROM @q;
      EXECUTE st;
      DEALLOCATE PREPARE st;
      SELECT UNIX_TIMESTAMP(@second_partition_date) INTO @second_partition_ts;
  
      SELECT CONCAT('Reorganizing first and second partitions. first partition date = ', @new_first_partition_date, ', second partition date = ', @second_partition_date);
   
      -- Then we build the ALTER TABLE query.
      SET @schema = p_schema;
      SET @table = p_table;
      SET @q = 'SELECT CONCAT(''ALTER TABLE '', @schema, ''.'', @table, '' REORGANIZE PARTITION p_first, '', @last_partition_name_to_keep, '' INTO ( PARTITION p_first VALUES LESS THAN ( '', @new_first_partition_ts, '' ), PARTITION '', @last_partition_name_to_keep, '' VALUES LESS THAN ( '', @second_partition_ts, '' ) ) '') INTO @query';
      PREPARE st FROM @q;
      EXECUTE st;
      DEALLOCATE PREPARE st;
     
      -- And then we prepare and execute the ALTER TABLE query.
      PREPARE st FROM @query;
      EXECUTE st;
      DEALLOCATE PREPARE st;
   END IF;
END$$
DELIMITER ;

Let’s try running the new procedure:

MariaDB [db1]> SHOW CREATE TABLE db1.quarterly_report_status\G
*************************** 1. row ***************************
       Table: quarterly_report_status
Create Table: CREATE TABLE `quarterly_report_status` (
  `report_id` int(11) NOT NULL,
  `report_status` varchar(20) NOT NULL,
  `report_updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1
/*!50100 PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated))
(PARTITION p_first VALUES LESS THAN (1475294400) ENGINE = InnoDB,
 PARTITION p201610 VALUES LESS THAN (1477972800) ENGINE = InnoDB,
 PARTITION p201611 VALUES LESS THAN (1480568400) ENGINE = InnoDB,
 PARTITION p201612 VALUES LESS THAN (1483246800) ENGINE = InnoDB,
 PARTITION p201701 VALUES LESS THAN (1485925200) ENGINE = InnoDB,
 PARTITION p201702 VALUES LESS THAN (1488344400) ENGINE = InnoDB,
 PARTITION p201703 VALUES LESS THAN (1491019200) ENGINE = InnoDB,
 PARTITION p201704 VALUES LESS THAN (1493611200) ENGINE = InnoDB,
 PARTITION p201705 VALUES LESS THAN (1496289600) ENGINE = InnoDB,
 PARTITION p201706 VALUES LESS THAN (1498881600) ENGINE = InnoDB,
 PARTITION p201707 VALUES LESS THAN (1501560000) ENGINE = InnoDB,
 PARTITION p201708 VALUES LESS THAN (1504238400) ENGINE = InnoDB,
 PARTITION p_future VALUES LESS THAN MAXVALUE ENGINE = InnoDB) */
1 row in set (0.00 sec)

MariaDB [db1]> CALL db1.drop_old_partitions('db1', 'quarterly_report_status', 6, 5);
+--------------------------------------------------------------------------+
| CONCAT('Dropping all partitions before: ', @last_partition_name_to_keep) |
+--------------------------------------------------------------------------+
| Dropping all partitions before: p201702                                  |
+--------------------------------------------------------------------------+
1 row in set (0.00 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201610                            |
+--------------------------------------------------------+
1 row in set (0.00 sec)

+---------------------------------------------------------+
| CONCAT('Sleeping for ', p_seconds_to_sleep, ' seconds') |
+---------------------------------------------------------+
| Sleeping for 5 seconds                                  |
+---------------------------------------------------------+
1 row in set (0.02 sec)

+---------------------------+
| SLEEP(p_seconds_to_sleep) |
+---------------------------+
|                         0 |
+---------------------------+
1 row in set (5.02 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201611                            |
+--------------------------------------------------------+
1 row in set (5.02 sec)

+---------------------------------------------------------+
| CONCAT('Sleeping for ', p_seconds_to_sleep, ' seconds') |
+---------------------------------------------------------+
| Sleeping for 5 seconds                                  |
+---------------------------------------------------------+
1 row in set (5.03 sec)

+---------------------------+
| SLEEP(p_seconds_to_sleep) |
+---------------------------+
|                         0 |
+---------------------------+
1 row in set (10.03 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201612                            |
+--------------------------------------------------------+
1 row in set (10.03 sec)

+---------------------------------------------------------+
| CONCAT('Sleeping for ', p_seconds_to_sleep, ' seconds') |
+---------------------------------------------------------+
| Sleeping for 5 seconds                                  |
+---------------------------------------------------------+
1 row in set (10.05 sec)

+---------------------------+
| SLEEP(p_seconds_to_sleep) |
+---------------------------+
|                         0 |
+---------------------------+
1 row in set (15.05 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201701                            |
+--------------------------------------------------------+
1 row in set (15.05 sec)

+--------------------------------------------------------------------------------------------------------------------------------------------------------------+
| CONCAT('Reorganizing first and second partitions. first partition date = ', @new_first_partition_date, ', second partition date = ', @second_partition_date) |
+--------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Reorganizing first and second partitions. first partition date = 2017-02-01 00:00:00, second partition date = 2017-03-01 00:00:00                            |
+--------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (15.06 sec)

Query OK, 0 rows affected (15.11 sec)

MariaDB [db1]> SHOW CREATE TABLE db1.quarterly_report_status\G
*************************** 1. row ***************************
       Table: quarterly_report_status
Create Table: CREATE TABLE `quarterly_report_status` (
  `report_id` int(11) NOT NULL,
  `report_status` varchar(20) NOT NULL,
  `report_updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1
/*!50100 PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated))
(PARTITION p_first VALUES LESS THAN (1485925200) ENGINE = InnoDB,
 PARTITION p201702 VALUES LESS THAN (1488344400) ENGINE = InnoDB,
 PARTITION p201703 VALUES LESS THAN (1491019200) ENGINE = InnoDB,
 PARTITION p201704 VALUES LESS THAN (1493611200) ENGINE = InnoDB,
 PARTITION p201705 VALUES LESS THAN (1496289600) ENGINE = InnoDB,
 PARTITION p201706 VALUES LESS THAN (1498881600) ENGINE = InnoDB,
 PARTITION p201707 VALUES LESS THAN (1501560000) ENGINE = InnoDB,
 PARTITION p201708 VALUES LESS THAN (1504238400) ENGINE = InnoDB,
 PARTITION p_future VALUES LESS THAN MAXVALUE ENGINE = InnoDB) */
1 row in set (0.00 sec)

We can see that our changes seem to be working as expected. In addition to old partitions being dropped, we can also see that p_first ‘s date range was updated.

Event definition

Our event definition only needs one minor change: the addition of the new parameter for our stored procedure. This is fairly straight forward. If we want it to sleep for 5 seconds between operations, then the new definition would be:

DROP EVENT db1.monthly_drop_old_partitions_event;

CREATE EVENT db1.monthly_drop_old_partitions_event
   ON SCHEDULE
   EVERY 1 MONTH
   STARTS NOW()
DO
   CALL db1.drop_old_partitions('db1', 'quarterly_report_status', 6, 5);

Conclusion

I enjoy the challenge of writing interesting stored procedures like this, so I will probably eventually write a stored procedure that will also automatically add new partitions. In the mean time, I would love to hear if anyone else can think of any other improvements.

Note: You can find part 3 of this blog series here .

Automatically Dropping Old Partitions in MySQL and MariaDB

Geoff Montee August 23, 2017 MariaDB , MySQL , Partitioning 10 Comments

A MariaDB Support customer recently asked how they could automatically drop old partitions after 6 months. MariaDB and MySQL do not have a mechanism to do this automatically out-of-the-box, but it is not too difficult to create a custom stored procedure and an event to call the procedure on the desired schedule. In this blog post, I will show one way to do that.

Partitioned table definition

For this demonstration, I’ll use a table definition based on one from MySQL’s documentation on range partitioning , with some minor changes:

DROP TABLE IF EXISTS db1.quarterly_report_status;

CREATE TABLE db1.quarterly_report_status (
   report_id INT NOT NULL,
   report_status VARCHAR(20) NOT NULL,
   report_updated TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB
PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated) ) (
   PARTITION p201610 VALUES LESS THAN ( UNIX_TIMESTAMP('2016-11-01 00:00:00')),
   PARTITION p201611 VALUES LESS THAN ( UNIX_TIMESTAMP('2016-12-01 00:00:00')),
   PARTITION p201612 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-01-01 00:00:00')),
   PARTITION p201701 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-02-01 00:00:00')),
   PARTITION p201702 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-03-01 00:00:00')),
   PARTITION p201703 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-04-01 00:00:00')),
   PARTITION p201704 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-05-01 00:00:00')),
   PARTITION p201705 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-06-01 00:00:00')),
   PARTITION p201706 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-07-01 00:00:00')),
   PARTITION p201707 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-08-01 00:00:00')),
   PARTITION p201708 VALUES LESS THAN ( UNIX_TIMESTAMP('2017-09-01 00:00:00')),
   PARTITION p_default VALUES LESS THAN (MAXVALUE)
);

The most significant change is that the partition naming scheme is based on the date. This will allow us to more easily determine which partitions to remove.

Stored procedure definition

The stored procedure itself contains some comments that explain what it does, so I will let the code speak for itself, for the most part:

DROP PROCEDURE IF EXISTS db1.drop_old_partitions;

DELIMITER $$
CREATE PROCEDURE db1.drop_old_partitions(p_schema varchar(64), p_table varchar(64), p_months_to_keep int)
   LANGUAGE SQL
   NOT DETERMINISTIC
   SQL SECURITY INVOKER
BEGIN  
   DECLARE done INT DEFAULT FALSE;
   DECLARE current_partition_name varchar(64);
   -- We'll use this cursor later to get
   -- the list of partitions to drop.
   -- @last_partition_name_to_keep will be
   -- set later.
   DECLARE cur1 CURSOR FOR 
   SELECT partition_name 
   FROM information_schema.partitions 
   WHERE TABLE_SCHEMA = p_schema 
   AND TABLE_NAME = p_table 
   AND PARTITION_NAME != 'p_default' 
   AND PARTITION_NAME < @last_partition_name_to_keep;
   DECLARE CONTINUE HANDLER FOR NOT FOUND SET done = TRUE;
   
   -- Now we get the last month of data that we want to keep
   -- by subtracting p_months_to_keep from the current date.
   -- Note that it will actually keep p_months_to_keep+1 partitions,
   -- since the current month is not complete.
   SET @date = CURDATE();
   SET @months_to_keep = p_months_to_keep;   
   SET @q = 'SELECT DATE_SUB(?, INTERVAL ? MONTH) INTO @last_month_to_keep';
   PREPARE st FROM @q;
   EXECUTE st USING @date, @months_to_keep;
   DEALLOCATE PREPARE st;
   
   -- Then we format the last month in the same format used
   -- in our partition names.
   SET @q = 'SELECT DATE_FORMAT(@last_month_to_keep, ''%Y%m'') INTO @formatted_last_month_to_keep';
   PREPARE st FROM @q;
   EXECUTE st;
   DEALLOCATE PREPARE st;
   
   -- And then we use the formatted date to build the name of the
   -- last partition that we want to keep. This partition name is
   -- assigned to @last_partition_name_to_keep, which is used in
   -- the cursor declared at the start of the procedure.
   SET @q = 'SELECT CONCAT(''p'', @formatted_last_month_to_keep) INTO @last_partition_name_to_keep';
   PREPARE st FROM @q;
   EXECUTE st;
   DEALLOCATE PREPARE st;
   
   SELECT CONCAT('Dropping all partitions before: ', @last_partition_name_to_keep);
   
   -- And then we loop through all partitions returned by the cursor,
   -- and those partitions are dropped.
   OPEN cur1;

   read_loop: LOOP
      FETCH cur1 INTO current_partition_name;
   
      IF done THEN
         LEAVE read_loop;
      END IF;

      SELECT CONCAT('Dropping partition: ', current_partition_name);
   
      -- First we build the ALTER TABLE query.
      SET @schema = p_schema;
      SET @table = p_table;
      SET @partition = current_partition_name;
      SET @q = 'SELECT CONCAT(''ALTER TABLE '', @schema, ''.'', @table, '' DROP PARTITION '', @partition) INTO @query';
      PREPARE st FROM @q;
      EXECUTE st;
      DEALLOCATE PREPARE st;
      
      -- And then we prepare and execute the ALTER TABLE query.
      PREPARE st FROM @query;
      EXECUTE st;
      DEALLOCATE PREPARE st;
   END LOOP;
END$$
DELIMITER ;

Let's try running the procedure as a test:

MariaDB [(none)]> SHOW CREATE TABLE db1.quarterly_report_status;
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table                   | Create Table                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| quarterly_report_status | CREATE TABLE `quarterly_report_status` (
  `report_id` int(11) NOT NULL,
  `report_status` varchar(20) NOT NULL,
  `report_updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1
/*!50100 PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated))
(PARTITION p201610 VALUES LESS THAN (1477972800) ENGINE = InnoDB,
 PARTITION p201611 VALUES LESS THAN (1480568400) ENGINE = InnoDB,
 PARTITION p201612 VALUES LESS THAN (1483246800) ENGINE = InnoDB,
 PARTITION p201701 VALUES LESS THAN (1485925200) ENGINE = InnoDB,
 PARTITION p201702 VALUES LESS THAN (1488344400) ENGINE = InnoDB,
 PARTITION p201703 VALUES LESS THAN (1491019200) ENGINE = InnoDB,
 PARTITION p201704 VALUES LESS THAN (1493611200) ENGINE = InnoDB,
 PARTITION p201705 VALUES LESS THAN (1496289600) ENGINE = InnoDB,
 PARTITION p201706 VALUES LESS THAN (1498881600) ENGINE = InnoDB,
 PARTITION p201707 VALUES LESS THAN (1501560000) ENGINE = InnoDB,
 PARTITION p201708 VALUES LESS THAN (1504238400) ENGINE = InnoDB,
 PARTITION p_default VALUES LESS THAN MAXVALUE ENGINE = InnoDB) */ |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

MariaDB [(none)]> CALL db1.drop_old_partitions('db1', 'quarterly_report_status', 6);
+--------------------------------------------------------------------------+
| CONCAT('Dropping all partitions before: ', @last_partition_name_to_keep) |
+--------------------------------------------------------------------------+
| Dropping all partitions before: p201702                                  |
+--------------------------------------------------------------------------+
1 row in set (0.00 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201610                            |
+--------------------------------------------------------+
1 row in set (0.00 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201611                            |
+--------------------------------------------------------+
1 row in set (0.01 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201612                            |
+--------------------------------------------------------+
1 row in set (0.04 sec)

+--------------------------------------------------------+
| CONCAT('Dropping partition: ', current_partition_name) |
+--------------------------------------------------------+
| Dropping partition: p201701                            |
+--------------------------------------------------------+
1 row in set (0.05 sec)

Query OK, 0 rows affected (0.07 sec)

MariaDB [(none)]> SHOW CREATE TABLE db1.quarterly_report_status;
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table                   | Create Table                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| quarterly_report_status | CREATE TABLE `quarterly_report_status` (
  `report_id` int(11) NOT NULL,
  `report_status` varchar(20) NOT NULL,
  `report_updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1
/*!50100 PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated))
(PARTITION p201702 VALUES LESS THAN (1488344400) ENGINE = InnoDB,
 PARTITION p201703 VALUES LESS THAN (1491019200) ENGINE = InnoDB,
 PARTITION p201704 VALUES LESS THAN (1493611200) ENGINE = InnoDB,
 PARTITION p201705 VALUES LESS THAN (1496289600) ENGINE = InnoDB,
 PARTITION p201706 VALUES LESS THAN (1498881600) ENGINE = InnoDB,
 PARTITION p201707 VALUES LESS THAN (1501560000) ENGINE = InnoDB,
 PARTITION p201708 VALUES LESS THAN (1504238400) ENGINE = InnoDB,
 PARTITION p_default VALUES LESS THAN MAXVALUE ENGINE = InnoDB) */ |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

As you can see from the above output, several partitions were dropped.

Event definition

We want our stored procedure to run automatically every month, so we can use an event to do that.
Before testing the event, we need to do two things:

We need to recreate the table with the original definition, so that it has all of the original partitions.
We need to ensure that event_scheduler=ON is set, and if not, we need to set it.

MariaDB [(none)]> SHOW GLOBAL VARIABLES LIKE 'event_scheduler';
+-----------------+-------+
| Variable_name   | Value |
+-----------------+-------+
| event_scheduler | OFF   |
+-----------------+-------+
1 row in set (0.00 sec)

MariaDB [(none)]> SET GLOBAL event_scheduler=ON;
Query OK, 0 rows affected (0.00 sec)

MariaDB [(none)]> SHOW GLOBAL VARIABLES LIKE 'event_scheduler';
+-----------------+-------+
| Variable_name   | Value |
+-----------------+-------+
| event_scheduler | ON    |
+-----------------+-------+
1 row in set (0.00 sec)

And then we can run the following:

CREATE EVENT db1.monthly_drop_old_partitions_event
   ON SCHEDULE
   EVERY 1 MONTH
   STARTS NOW()
DO
   CALL db1.drop_old_partitions('db1', 'quarterly_report_status', 6);

Since we specified STARTS NOW() , the event ran immediately. We can confirm this by looking at the table definition again:

MariaDB [(none)]> SHOW CREATE TABLE db1.quarterly_report_status;
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table                   | Create Table                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| quarterly_report_status | CREATE TABLE `quarterly_report_status` (
  `report_id` int(11) NOT NULL,
  `report_status` varchar(20) NOT NULL,
  `report_updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1
/*!50100 PARTITION BY RANGE ( UNIX_TIMESTAMP(report_updated))
(PARTITION p201702 VALUES LESS THAN (1488344400) ENGINE = InnoDB,
 PARTITION p201703 VALUES LESS THAN (1491019200) ENGINE = InnoDB,
 PARTITION p201704 VALUES LESS THAN (1493611200) ENGINE = InnoDB,
 PARTITION p201705 VALUES LESS THAN (1496289600) ENGINE = InnoDB,
 PARTITION p201706 VALUES LESS THAN (1498881600) ENGINE = InnoDB,
 PARTITION p201707 VALUES LESS THAN (1501560000) ENGINE = InnoDB,
 PARTITION p201708 VALUES LESS THAN (1504238400) ENGINE = InnoDB,
 PARTITION p_default VALUES LESS THAN MAXVALUE ENGINE = InnoDB) */ |
+-------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

Conclusion

Thanks to the flexibility of stored procedures and events, it is relatively easy to automatically drop old partitions in MySQL and MariaDB. Has anyone else implemented something like this?

Note: You can find part 2 of this blog series here .

Using the MariaDB Audit Plugin with MySQL

Geoff Montee December 5, 2016 MariaDB , MySQL , Security 8 Comments

The MariaDB audit plugin is an audit plugin that is bundled with MariaDB server. However, even though it is bundled with MariaDB, the plugin is actually compatible with MySQL as well. In this blog post, I will describe how to install the plugin with MySQL.

Install the plugin

Unfortunately, neither MariaDB Corporation nor MariaDB Foundation currently distribute a standalone binary for the MariaDB audit plugin. That means that if you want to use this plugin with MySQL, you will have to obtain the plugin from a MariaDB server package. We can check this table to determine what version of MariaDB server that we should use. The table says that the latest version of the plugin is 1.4.0, and that this version is present in MariaDB 10.1.11. The latest release of MariaDB 10.1 is currently 10.1.19, so let’s just grab that, since that should also have the plugin:

$ wget https://downloads.mariadb.org/interstitial/mariadb-10.1.19/bintar-linux-x86_64/mariadb-10.1.19-linux-x86_64.tar.gz

Let’s extract the tarball and copy the plugin library from the tarball’s plugin directory to MySQL’s plugin directory:

$ tar -xzf mariadb-10.1.19-linux-x86_64.tar.gz $ ls -l mariadb-10.1.19-linux-x86_64/lib/plugin/ | grep "audit" -rwxr-xr-x 1 ec2-user ec2-user 176024 Nov 4 09:37 server_audit.so $ sudo install mariadb-10.1.19-linux-x86_64/lib/plugin/server_audit.so /usr/lib64/mysql/plugin/

Now that the plugin library is in MySQL’s plugin directory, we can tell MySQL to install it:

$ mysql -u root Welcome to the MySQL monitor. Commands end with ; or \g. Your MySQL connection id is 3 Server version: 5.6.30-log MySQL Community Server (GPL)

Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

mysql> INSTALL PLUGIN server_audit SONAME 'server_audit.so';
Query OK, 0 rows affected (0.02 sec)

Configure the plugin

Now that the plugin is installed, we can configure it. For example, if we want to log all 6 event types, but we want to exclude the user named root , then we could add the following to MySQL’s configuration file:

server_audit_logging=ON server_audit_events=connect,query,table,query_ddl,query_dml,query_dcl server_audit_excl_users=root

And then restart the server:

$ sudo systemctl restart mysqld

At that point, audit logging will be enabled!

For more information on configuring MariaDB’s audit plugin, see this documentation page .

Has anyone used the MariaDB audit plugin with MySQL?

Importing InnoDB Partitions in MySQL 5.6 and MariaDB 10.0/10.1

Geoff Montee December 5, 2016 MariaDB , MySQL 4 Comments

Transportable tablespaces for InnoDB tables is a very useful feature added in MySQL 5.6 and MariaDB 10.0. With this new feature, an InnoDB table’s tablespace file can be copied from one server to another, as long as the table uses a file-per-table tablespace .

Unfortunately, the initial transportable tablespace feature in MySQL 5.6 and MariaDB 10.0 does not support partitioned tables. Support for partitioned tables was added in MySQL 5.7 . MDEV-10568 has been submitted to request that this feature be ported to MariaDB, so hopefully that does happen at some point. However, having this feature in some future new version doesn’t help you much if you wanted to use this feature in the older versions of MySQL or MariaDB.

The good news is that there is a workaround that allows you to use transportable tablespaces in MySQL 5.6 and MariaDB 10.0/10.1/10.2 to copy partitioned tables from one server to another. In this blog post, I will describe how to do so. This process can be a bit tedious, so I would recommend writing a script to automate it.

Test data

In this post, I’ll use the following test table to demonstrate how this works:

CREATE TABLE employees ( id INT NOT NULL, fname VARCHAR(30), lname VARCHAR(30), store_id INT NOT NULL ) PARTITION BY RANGE (store_id) ( PARTITION p0 VALUES LESS THAN (6), PARTITION p1 VALUES LESS THAN (11), PARTITION p2 VALUES LESS THAN (16), PARTITION p3 VALUES LESS THAN MAXVALUE ); INSERT INTO employees VALUES (1, 'Geoff', 'Montee', 1), (2, 'Chris', 'Calendar', 6), (3, 'Kyle', 'Joiner', 11), (4, 'Will', 'Fong', 16);

Export table files from original server

The process to export the partitioned table’s tablespaces from the original server is almost identical to the process for non-partitioned tables.

The first step, is to execute the following FLUSH command on the table:

MariaDB [db1]> FLUSH TABLES employees FOR EXPORT; Query OK, 0 rows affected (0.00 sec)

After executing the above command, leave the session open, so that the tables are locked.

Next, you should see some .ibd and .cfg files for the table in the database’s data directory:

$ sudo ls -l /var/lib/mysql/db1/ total 428 -rw-rw---- 1 mysql mysql 827 Dec 5 16:08 employees.frm -rw-rw---- 1 mysql mysql 48 Dec 5 16:08 employees.par -rw-rw---- 1 mysql mysql 579 Dec 5 18:47 employees#P#p0.cfg -rw-r----- 1 mysql mysql 98304 Dec 5 16:43 employees#P#p0.ibd -rw-rw---- 1 mysql mysql 579 Dec 5 18:47 employees#P#p1.cfg -rw-rw---- 1 mysql mysql 98304 Dec 5 16:08 employees#P#p1.ibd -rw-rw---- 1 mysql mysql 579 Dec 5 18:47 employees#P#p2.cfg -rw-rw---- 1 mysql mysql 98304 Dec 5 16:08 employees#P#p2.ibd -rw-rw---- 1 mysql mysql 579 Dec 5 18:47 employees#P#p3.cfg -rw-rw---- 1 mysql mysql 98304 Dec 5 16:08 employees#P#p3.ibd

Copy these files somewhere safe:

$ mkdir /tmp/backup $ cp /var/lib/mysql/db1/employees*ibd /tmp/backup/ $ cp /var/lib/mysql/db1/employees*cfg /tmp/backup/ $ ls -l /tmp/backup/ total 400 -rw-r----- 1 root root 579 Dec 5 18:52 employees#P#p0.cfg -rw-r----- 1 root root 98304 Dec 5 18:52 employees#P#p0.ibd -rw-r----- 1 root root 579 Dec 5 18:52 employees#P#p1.cfg -rw-r----- 1 root root 98304 Dec 5 18:52 employees#P#p1.ibd -rw-r----- 1 root root 579 Dec 5 18:52 employees#P#p2.cfg -rw-r----- 1 root root 98304 Dec 5 18:52 employees#P#p2.ibd -rw-r----- 1 root root 579 Dec 5 18:52 employees#P#p3.cfg -rw-r----- 1 root root 98304 Dec 5 18:52 employees#P#p3.ibd

Now that the files are copied, you can unlock the tables in the session that you still have open:

MariaDB [db1]> UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec)

Import table files on new server

Now that we have the .ibd and .cfg files of the partitions, the first step would be to place them somewhere where they will be accessible on your new server.

Then, if it does not already exist, create an empty copy of the partitioned table:

MariaDB [newdb]> CREATE TABLE employees ( -> id INT NOT NULL, -> fname VARCHAR(30), -> lname VARCHAR(30), -> store_id INT NOT NULL -> ) -> PARTITION BY RANGE (store_id) ( -> PARTITION p0 VALUES LESS THAN (6), -> PARTITION p1 VALUES LESS THAN (11), -> PARTITION p2 VALUES LESS THAN (16), -> PARTITION p3 VALUES LESS THAN MAXVALUE -> ); Query OK, 0 rows affected (0.06 sec)

Now we need an empty non-partitioned table that has the same structure as our partitioned table to serve as a placeholder. We can create that with the following query:

MariaDB [newdb]> CREATE TABLE placeholder AS SELECT * FROM employees WHERE NULL; Query OK, 0 rows affected (0.01 sec) Records: 0 Duplicates: 0 Warnings: 0

The above query gets us a non-partitioned table with the original structure that has 0 rows:

MariaDB [newdb]> SHOW CREATE TABLE placeholder; +-------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Table | Create Table | +-------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | placeholder | CREATE TABLE `placeholder` ( `id` int(11) NOT NULL, `fname` varchar(30) DEFAULT NULL, `lname` varchar(30) DEFAULT NULL, `store_id` int(11) NOT NULL ) ENGINE=InnoDB DEFAULT CHARSET=latin1 | +-------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec)

MariaDB [newdb]> SELECT * FROM placeholder;
Empty set (0.00 sec)

After this point is where the process can get a little tedious if your table has a lot of partitions. For each partition, we need to do the following:

Discard our placeholder table’s tablespace:

MariaDB [newdb]> ALTER TABLE placeholder DISCARD TABLESPACE; Query OK, 0 rows affected (0.00 sec)

Copy the .ibd and .cfg files for the partition to the database’s data directory, but rename these files such that they are named for the placeholder table:

$ cp /tmp/backup/employees#P#p0.cfg /var/lib/mysql/newdb/placeholder.cfg $ cp /tmp/backup/employees#P#p0.ibd /var/lib/mysql/newdb/placeholder.ibd $ chown mysql:mysql /var/lib/mysql/newdb/placeholder.*

Import the tablespace for the placeholder table:

MariaDB [newdb]> ALTER TABLE placeholder IMPORT TABLESPACE; Query OK, 0 rows affected (0.04 sec)

The placeholder table now contains the data of p0 from the original partitioned table:

Now exchange partition p0 in our partitioned table with the tablespace of our placeholder table:

MariaDB [newdb]> ALTER TABLE employees EXCHANGE PARTITION p0 WITH TABLE placeholder; Query OK, 0 rows affected (0.02 sec)

Now our partitioned table on the new server has the real contents of partition p0 :

If we repeat the above process for partitions p1 , p2 , and p3 , then our partitioned table on the new server will have all of the contents of the table from the original server:

MariaDB [newdb]> SELECT * FROM employees; +----+-------+----------+----------+ | id | fname | lname | store_id | +----+-------+----------+----------+ | 1 | Geoff | Montee | 1 | | 2 | Chris | Calendar | 6 | | 3 | Kyle | Joiner | 11 | | 4 | Will | Fong | 16 | +----+-------+----------+----------+ 4 rows in set (0.00 sec)

Has anyone successfully used a process like this in the past?

Emulating Sequences in MySQL and MariaDB

Geoff Montee August 18, 2016 MariaDB , MySQL 8 Comments

Sequences are objects defined by the SQL standard that are used to create monotonically increasing sequences of numeric values. Whenever nextval is called on a sequence object, it generates and returns the next number in the sequence. For MySQL and MariaDB users, this might sound similar to MySQL’s AUTO_INCREMENT columns , but there are some differences:

Sequences are defined by the SQL Standard. AUTO_INCREMENT columns are not in the standard, but are a MySQL extension.
Sequences are their own objects with their own state, which means that multiple columns in multiple tables could all use numbers from the same sequence. In contrast, MySQL’s AUTO_INCREMENT feature is tied to a specific column in a specific table, so multiple columns in multiple tables cannot directly use the same AUTO_INCREMENT pool.

MySQL and MariaDB do not yet support SQL Standard sequences. If you would like MariaDB to support sequences, you may want to consider voting for this feature request .

Users who have migrated to MySQL or MariaDB from other databases might find this feature to be a strange omission, considering that many other databases do support sequences, including:

Despite the fact that MySQL and MariaDB don’t yet support sequences, it is fairly easy to emulate SQL standard sequences in MySQL and MariaDB using an AUTO_INCREMENT column and functions. In this blog post, I’ll describe how to do that using MariaDB 10.1.

Emulating sequences in MariaDB

The first step needed to create our emulated sequence is to create a table that keeps track of the sequence values:

CREATE TABLE sequence_values ( id INT AUTO_INCREMENT PRIMARY KEY, thread_id INT NOT NULL, created DATETIME DEFAULT CURRENT_TIMESTAMP );

The second step is to create a function that generates and returns the next value in the sequence:

DELIMITER //

CREATE FUNCTION `sequence_nextval`()
RETURNS INT
NOT DETERMINISTIC
MODIFIES SQL DATA
BEGIN

DECLARE nextval int;

INSERT INTO sequence_values (thread_id) VALUES (CONNECTION_ID());
SELECT last_insert_id() INTO nextval;

RETURN nextval;

END//

DELIMITER ;

Finally, let’s create a table that we want to use the sequence with:

CREATE TABLE sequence_test_a ( seq int NOT NULL PRIMARY KEY, str varchar(50) );

For users who are used to databases with real standard sequence support, it might be tempting to define the table in the following way instead:

CREATE TABLE sequence_test_a ( seq int NOT NULL PRIMARY KEY DEFAULT sequence_nextval(), str varchar(50) );

Unfortunately, MariaDB 10.1 does not support setting a DEFAULT value to a stored function. ~~However, this will be supported in MariaDB 10.2~~ . ( Edit: I’ve been told that new MariaDB 10.2 feature will not include support for stored functions as DEFAULT values. However, you can still use triggers, as mentioned in the comments by Frederico.)

One of the benefits of sequences is that they can be used across multiple tables, so let’s create a second table that will use the sequence as well:

CREATE TABLE sequence_test_b ( seq int NOT NULL PRIMARY KEY, str varchar(50) );

Now let’s insert some data into the tables:

MariaDB [db1]> INSERT INTO sequence_test_a VALUES (sequence_nextval(), 'a_str1'); Query OK, 1 row affected (0.00 sec)

MariaDB [db1]> INSERT INTO sequence_test_a VALUES (sequence_nextval(), 'a_str2');
Query OK, 1 row affected (0.01 sec)

MariaDB [db1]> INSERT INTO sequence_test_b VALUES (sequence_nextval(), 'b_str1');
Query OK, 1 row affected (0.00 sec)

MariaDB [db1]> INSERT INTO sequence_test_b VALUES (sequence_nextval(), 'b_str2');
Query OK, 1 row affected (0.00 sec)

MariaDB [db1]> INSERT INTO sequence_test_a VALUES (sequence_nextval(), 'a_str3');
Query OK, 1 row affected (0.00 sec)

What are the contents of these tables now?

MariaDB [db1]> SELECT * FROM sequence_test_a; +-----+--------+ | seq | str | +-----+--------+ | 1 | a_str1 | | 2 | a_str2 | | 5 | a_str3 | +-----+--------+ 3 rows in set (0.00 sec)

MariaDB [db1]> SELECT * FROM sequence_test_b;
+-----+--------+
| seq | str |
+-----+--------+
| 3 | b_str1 |
| 4 | b_str2 |
+-----+--------+
2 rows in set (0.00 sec)

As you can see from the above output, the seq column in each table was populated with monotonically increasing values in the order in which the rows were inserted, so our sequence appears to be working properly.

I should also note that the sequence_values table will grow over time:

MariaDB [db1]> SELECT * FROM sequence_values; +----+-----------+---------------------+ | id | thread_id | created | +----+-----------+---------------------+ | 1 | 3 | 2016-08-18 14:09:49 | | 2 | 3 | 2016-08-18 14:09:50 | | 3 | 3 | 2016-08-18 14:09:58 | | 4 | 3 | 2016-08-18 14:10:22 | | 5 | 3 | 2016-08-18 14:10:23 | +----+-----------+---------------------+ 5 rows in set (0.00 sec)

If you do not need to keep track of when a sequence was generated, you could create an event or cron job to periodically prune old events.

Has anyone else created their own sequence implementation in MySQL or MariaDB?

Bitwise operators with BINARY fields in MySQL and MariaDB

Geoff Montee August 10, 2016 MariaDB , MySQL Leave a Comment

A MariaDB support customer recently upgraded to MariaDB 10.1, and they noticed that some of their queries using bitwise operators started to return warnings, which they thought was strange because they produced no warnings in MariaDB 10.0. These particular queries used bitwise operators on BINARY(N) fields.

For example, their table was similar to this:

CREATE TABLE item_flags ( item_id int(11) NOT NULL, flags binary(2) NOT NULL DEFAULT '\0\0', PRIMARY KEY (`item_id`) );

And their query was similar to this:

SELECT item_id, flags FROM item_flags WHERE (flags & 4) = 4;

Let’s see what happens when we actually execute this query:

MariaDB [db1]> CREATE TABLE item_flags ( -> item_id int(11) NOT NULL, -> flags binary(2) NOT NULL DEFAULT '\0\0', -> PRIMARY KEY (`item_id`) -> ); Query OK, 0 rows affected (0.01 sec)

MariaDB [db1]> INSERT INTO item_flags VALUES (1, 1), (2, 2), (3, 3), (4, 4);
Query OK, 4 rows affected (0.01 sec)
Records: 4 Duplicates: 0 Warnings: 0

MariaDB [db1]> SELECT item_id, flags
-> FROM item_flags
-> WHERE (flags & 4) = 4;
+---------+-------+
| item_id | flags |
+---------+-------+
| 4 | 4 |
+---------+-------+
1 row in set, 4 warnings (0.00 sec)

As we can see from the above output, it looks like MariaDB gave us a warning for each row that the query examined. Let’s look at those warnings:

The warnings show us two things:

When each row was inserted, the flags column was treated as a binary string that was right-padded with the null character. From the MySQL documentation :

The BINARY and VARBINARY types are similar to CHAR and VARCHAR, except that they contain binary strings rather than nonbinary strings. That is, they contain byte strings rather than character strings. This means that they have no character set, and sorting and comparison are based on the numeric values of the bytes in the values.

…snip…

When BINARY values are stored, they are right-padded with the pad value to the specified length. The pad value is 0x00 (the zero byte). Values are right-padded with 0x00 on insert, and no trailing bytes are removed on select. All bytes are significant in comparisons, including ORDER BY and DISTINCT operations. 0x00 bytes and spaces are different in comparisons, with 0x00 < space.
That string value is then being converted to an INTEGER . This is because bitwise operators in MySQL and MariaDB operate on integers. From the MySQL documentation :

Bit functions and operators comprise BIT_COUNT(), BIT_AND(), BIT_OR(), BIT_XOR(), &, |, ^, ~, <>. (The BIT_AND(), BIT_OR(), and BIT_XOR() functions are aggregate functions described at Section 13.20.1, “Aggregate (GROUP BY) Function Descriptions”.) Currently, bit functions and operators require BIGINT (64-bit integer) arguments and return BIGINT values, so they have a maximum range of 64 bits. Arguments of other types are converted to BIGINT and truncation might occur.

It would probably be an improvement if the flags column’s data type were tinyint instead of BINARY(2) , so that this conversion step could be avoided.

However, it sounds like bitwise operators will work directly on BINARY(N) fields in MySQL 8.0. From the MySQL documentation again:

A planned extension for MySQL 8.0 is to change this cast-to-BIGINT behavior: Bit functions and operators will permit binary string type arguments (BINARY, VARBINARY, and the BLOB types), enabling them to take arguments and produce return values larger than 64 bits. Consequently, bit operations on binary arguments in MySQL 5.7 might produce different results in MySQL 8.0. To provide advance notice about this potential change in behavior, the server produces warnings as of MySQL 5.7.11 for bit operations for which binary arguments will not be converted to integer in MySQL 8.0. These warnings afford an opportunity to rewrite affected statements. To explicitly produce MySQL 5.7 behavior in a way that will not change after an upgrade to 8.0, cast bit-operation binary arguments to convert them to integer.

This sounds like it would be a nice improvement. In MySQL 8.0, it sounds like the above SQL could be rewritten like this:

MariaDB [db1]> CREATE TABLE item_flags ( -> item_id int(11) NOT NULL, -> flags binary(2) NOT NULL DEFAULT x'00', -> PRIMARY KEY (`item_id`) -> ); Query OK, 0 rows affected (0.01 sec)

MariaDB [db1]> INSERT INTO item_flags VALUES (1, x'01'), (2, x'02'), (3, x'03'), (4, x'04');
Query OK, 4 rows affected (0.00 sec)
Records: 4 Duplicates: 0 Warnings: 0

MariaDB [db1]> SELECT item_id, flags
-> FROM item_flags
-> WHERE (flags & x'04') = x'04';
+---------+-------+
| item_id | flags |
+---------+-------+
| 1 | |
| 2 | |
| 3 | |
| 4 | |
+---------+-------+
4 rows in set, 9 warnings (0.00 sec)

If it isn’t completely obvious from the above output, this SQL doesn’t currently work the way some might expect it to in MariaDB 10.1, since the BINARY(N) field has to be converted to bigint to make use of the bitwise-and ( & ) operator.

I submitted a feature request to have BINARY(N) support for bitwise operators implemented in MariaDB . If this feature sounds important to you, you may want to consider voting for it to express your interest.

DDL Failures in MariaDB Galera Cluster

Geoff Montee July 28, 2016 DDL , Galera Cluster , MariaDB , MySQL Leave a Comment

A MariaDB support customer recently asked me what would happen if a Data Definition Language (DDL) statement failed to complete on one or more nodes in MariaDB Galera Cluster. In this blog post, I will demonstrate what would happen.

The demonstration below was performed on a 2-node cluster running MariaDB 10.1 , but other Galera Cluster distributions should work similarly.

Schema Upgrades in Galera Cluster

Schema upgrades and DDL in Galera Cluster are handled a bit differently than in a standalone MariaDB or MySQL server.

Transactions in Galera Cluster are replicated in a “virtually synchronous” manner. This means that unless a particular node is desynchronized from the cluster, all replicated tables need to have identical (or at least compatible) definitions on all nodes. If a node tries to replicate data for a particular table and if some nodes have incompatible definitions for that table, those nodes will not be able to apply the transactions to their copy of the table. This also means that incompatible schema upgrades should happen on all nodes at the same time.

Galera Cluster provides two methods of applying schema upgrades, and you can switch between them using the wsrep_OSU_method option. One method, Total Order Isolation (TOI), can be used to apply incompatible changes in a slow, but safe way. The other method, Rolling Schema Upgrade (RSU), can be used to apply backward-compatible changes in a faster way. These are described in more detail in the Galera Cluster documentation page about Schema Upgrades .

But since DDL is treated specially in Galera Cluster, what happens if some DDL fails to complete successfully on one or more nodes?

A DDL Failure in TOI Mode

First, lets look at what happens when DDL fails in TOI mode. Let’s make sure that TOI mode is currently set:

It is, so let’s create a table by executing the following on one node:

MariaDB [db1]> CREATE TABLE tab ( -> id int PRIMARY KEY, -> str varchar(50) -> ) ENGINE=InnoDB; Query OK, 0 rows affected (0.03 sec)

Let’s make sure that this table exists on both nodes.

Node 1:

MariaDB [db1]> SHOW CREATE TABLE tab; +-------+---------------------------------------------------------------------------------------------------------------------------------------------+ | Table | Create Table | +-------+---------------------------------------------------------------------------------------------------------------------------------------------+ | tab | CREATE TABLE `tab` ( `id` int(11) NOT NULL, `str` varchar(50) DEFAULT NULL, PRIMARY KEY (`id`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1 | +-------+---------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec)

Node 2:

We want to see DDL fail, so lets do some setup for that by making one node have a slightly different definition of the table. We can do so by running some DDL in RSU mode:

MariaDB [db1]> SET wsrep_osu_method='RSU'; Query OK, 0 rows affected (0.00 sec)

MariaDB [db1]> ALTER TABLE tab ADD COLUMN num int DEFAULT NULL;
Query OK, 0 rows affected (0.02 sec)
Records: 0 Duplicates: 0 Warnings: 0

Do the two nodes have the same definition of the table now?

Node 1:

MariaDB [db1]> SHOW CREATE TABLE tab; +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Table | Create Table | +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | tab | CREATE TABLE `tab` ( `id` int(11) NOT NULL, `str` varchar(50) DEFAULT NULL, `num` int(11) DEFAULT NULL, PRIMARY KEY (`id`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1 | +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec)

Node 2:

They now have different definitions, so let’s execute some DDL on node 1 that will fail on node 2. We also need to set wsrep_OSU_method back to TOI.

MariaDB [db1]> SET wsrep_osu_method='TOI'; Query OK, 0 rows affected (0.00 sec)

MariaDB [db1]> ALTER TABLE tab MODIFY COLUMN num bigint DEFAULT NULL;
Query OK, 0 rows affected (0.04 sec)
Records: 0 Duplicates: 0 Warnings: 0

Since node 2 does not have the num column, this DDL should fail on that node. Lets look at the definition of the table on both nodes now:

Node 1:

MariaDB [db1]> SHOW CREATE TABLE tab; +-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Table | Create Table | +-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | tab | CREATE TABLE `tab` ( `id` int(11) NOT NULL, `str` varchar(50) DEFAULT NULL, `num` bigint(20) DEFAULT NULL, PRIMARY KEY (`id`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1 | +-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ 1 row in set (0.00 sec)

Node 2:

The DDL obviously failed on node 2, but it doesn’t look like anything happened. Lets look at node 2’s error log:

2016-07-28 14:42:48 140579533392640 [ERROR] Slave SQL: Error ‘Unknown column ‘num’ in ‘tab” on query. Default database: ‘db1’. Query: ‘ALTER TABLE tab MODIFY COLUMN num bigint DEFAULT NULL’, Internal MariaDB error code: 1054
2016-07-28 14:42:48 140579533392640 [Warning] WSREP: RBR event 1 Query apply warning: 1, 3
2016-07-28 14:42:48 140579533392640 [Warning] WSREP: Ignoring error for TO isolated action: source: d25d604b-54f0-11e6-a77e-f681b74c50f4 version: 3 local: 0 state: APPLYING flags: 65 conn_id: 5 trx_id: -1 seqnos (l: 7, g: 3, s: 2, d: 2, ts: 1017404963701)

Node 2 just ignored the error!

Now what actually happens when we try to insert something into the num field?

Node 1:

MariaDB [db1]> INSERT INTO tab (id, str, num) VALUES (1, 'str1', 1); Query OK, 1 row affected (0.01 sec)

MariaDB [db1]> SELECT * FROM db1.tab;
+----+------+------+
| id | str | num |
+----+------+------+
| 1 | str1 | 1 |
+----+------+------+
1 row in set (0.00 sec)

Node 2:

MariaDB [db1]> SELECT * FROM tab; +----+------+ | id | str | +----+------+ | 1 | str1 | +----+------+ 1 row in set (0.00 sec)

The extra column at the end of the list is just ignored! This is because Galera Cluster follows many of the same compatibility rules as standard MySQL replication, and an extra column at the end of the list is considered a valid difference in standard MySQL replication .

But lets see what happens if the difference is invalid.

Node 1:

MariaDB [db1]> SET wsrep_osu_method='RSU'; Query OK, 0 rows affected (0.00 sec)

MariaDB [db1]> ALTER TABLE tab DROP COLUMN num;
Query OK, 0 rows affected (0.02 sec)
Records: 0 Duplicates: 0 Warnings: 0

MariaDB [db1]> ALTER TABLE tab ADD COLUMN num int DEFAULT NULL AFTER id;
Query OK, 0 rows affected (0.03 sec)
Records: 0 Duplicates: 0 Warnings: 0

MariaDB [db1]> SHOW CREATE TABLE tab;
+-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table | Create Table |
+-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| tab | CREATE TABLE `tab` (
`id` int(11) NOT NULL,
`num` int(11) DEFAULT NULL,
`str` varchar(50) DEFAULT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 |
+-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

Node 2:

And now let’s try to insert some data:

Node 1:

MariaDB [db1]> INSERT INTO tab (id, num, str) VALUES (2, 1, 'str2'); Query OK, 1 row affected (0.00 sec)

MariaDB [db1]> SELECT * FROM db1.tab;
+----+------+------+
| id | num | str |
+----+------+------+
| 1 | NULL | str1 |
| 2 | 1 | str2 |
+----+------+------+
2 rows in set (0.00 sec)

Node 2:

MariaDB [db1]> SELECT * FROM tab; ERROR 2006 (HY000): MySQL server has gone away No connection. Trying to reconnect... Connection id: 5 Current database: db1

+----+------+------+
| id | num | str |
+----+------+------+
| 1 | NULL | str1 |
| 2 | 1 | str2 |
+----+------+------+
2 rows in set (0.00 sec)

You might notice two weird things here:

Our client was disconnected from node 2.
Node 2 has the num column now.

That’s weird! Let’s look at node 2’s error log. What do we see?

First, we can see that it tried to apply the transaction 4 times:

2016-07-28 14:55:34 140579533392640 [ERROR] Slave SQL: Column 1 of table ‘db1.tab’ cannot be converted from type ‘int’ to type ‘varchar(50)’, Internal MariaDB error code: 1677
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: RBR event 2 Write_rows_v1 apply warning: 3, 5
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: Failed to apply app buffer: seqno: 5, status: 1
at galera/src/trx_handle.cpp:apply():351
Retrying 2th time
2016-07-28 14:55:34 140579533392640 [ERROR] Slave SQL: Column 1 of table ‘db1.tab’ cannot be converted from type ‘int’ to type ‘varchar(50)’, Internal MariaDB error code: 1677
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: RBR event 2 Write_rows_v1 apply warning: 3, 5
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: Failed to apply app buffer: seqno: 5, status: 1
at galera/src/trx_handle.cpp:apply():351
Retrying 3th time
2016-07-28 14:55:34 140579533392640 [ERROR] Slave SQL: Column 1 of table ‘db1.tab’ cannot be converted from type ‘int’ to type ‘varchar(50)’, Internal MariaDB error code: 1677
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: RBR event 2 Write_rows_v1 apply warning: 3, 5
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: Failed to apply app buffer: seqno: 5, status: 1
at galera/src/trx_handle.cpp:apply():351
Retrying 4th time
2016-07-28 14:55:34 140579533392640 [ERROR] Slave SQL: Column 1 of table ‘db1.tab’ cannot be converted from type ‘int’ to type ‘varchar(50)’, Internal MariaDB error code: 1677
2016-07-28 14:55:34 140579533392640 [Warning] WSREP: RBR event 2 Write_rows_v1 apply warning: 3, 5
2016-07-28 14:55:34 140579533392640 [ERROR] WSREP: Failed to apply trx: source: d25d604b-54f0-11e6-a77e-f681b74c50f4 version: 3 local: 0 state: APPLYING flags: 1 conn_id: 5 trx_id: 76646 seqnos (l: 9, g: 5, s: 4, d: 3, ts: 1783539089510)

When that failed, the failed node determined that it was inconsistent with the cluster, so it shot itself in the head :

2016-07-28 14:55:34 140579533392640 [ERROR] WSREP: Failed to apply trx 5 4 times
2016-07-28 14:55:34 140579533392640 [ERROR] WSREP: Node consistency compromized, aborting…
2016-07-28 14:55:34 140579533392640 [Note] WSREP: Closing send monitor…
2016-07-28 14:55:34 140579533392640 [Note] WSREP: Closed send monitor.
2016-07-28 14:55:34 140579533392640 [Note] WSREP: gcomm: terminating thread
2016-07-28 14:55:34 140579533392640 [Note] WSREP: gcomm: joining thread
2016-07-28 14:55:34 140579533392640 [Note] WSREP: gcomm: closing backend
…snip…
2016-07-28 14:55:35 140579533392640 [Note] WSREP: /usr/sbin/mysqld: Terminated.

And it was automatically restarted by systemd, at which point it did an SST:

2016-07-28 14:55:44 139821320411264 [Note] WSREP: Read nil XID from storage engines, skipping position init
2016-07-28 14:55:44 139821320411264 [Note] WSREP: wsrep_load(): loading provider library ‘/usr/lib64/galera/libgalera_smm.so’
2016-07-28 14:55:44 139821320411264 [Note] WSREP: wsrep_load(): Galera 25.3.15(r3578) by Codership Oy loaded successfully.
2016-07-28 14:55:44 139821320411264 [Note] WSREP: CRC-32C: using hardware acceleration.
2016-07-28 14:55:44 139821320411264 [Note] WSREP: Found saved state: 00000000-0000-0000-0000-000000000000:-1
…snip…
2016-07-28 14:55:45 139821320096512 [Note] WSREP: New cluster view: global state: d25dbeb7-54f0-11e6-bac9-c2bc3c331fb6:5, view# 4: Primary, number of nodes: 2, my index: 1, protocol version 3
2016-07-28 14:55:45 139821320096512 [Warning] WSREP: Gap in state sequence. Need state transfer.
2016-07-28 14:55:45 139821024540416 [Note] WSREP: Running: ‘wsrep_sst_rsync –role ‘joiner’ –address ‘172.31.22.174’ –datadir ‘/var/lib/mysql/’ –parent ‘2159’ –binlog ‘mariadb-bin’ ‘
2016-07-28 14:55:45 139821320096512 [Note] WSREP: Prepared SST request: rsync|172.31.22.174:4444/rsync_sst
2016-07-28 14:55:45 139821320096512 [Note] WSREP: wsrep_notify_cmd is not defined, skipping notification.
2016-07-28 14:55:45 139821320096512 [Note] WSREP: REPL Protocols: 7 (3, 2)
2016-07-28 14:55:45 139821097465600 [Note] WSREP: Service thread queue flushed.
2016-07-28 14:55:45 139821320096512 [Note] WSREP: Assign initial position for certification: 5, protocol version: 3
2016-07-28 14:55:45 139821097465600 [Note] WSREP: Service thread queue flushed.
2016-07-28 14:55:45 139821320096512 [Warning] WSREP: Failed to prepare for incremental state transfer: Local state UUID (00000000-0000-0000-0000-000000000000) does not match group state UUID (d25dbeb7-54f0-11e6-bac9-c2bc3c331fb6): 1 (Operation not permitted)
at galera/src/replicator_str.cpp:prepare_for_IST():482. IST will be unavailable.
2016-07-28 14:55:45 139821041313536 [Note] WSREP: Member 1.0 () requested state transfer from ‘*any*’. Selected 0.0 ()(SYNCED) as donor.
2016-07-28 14:55:45 139821041313536 [Note] WSREP: Shifting PRIMARY -> JOINER (TO: 5)
2016-07-28 14:55:45 139821320096512 [Note] WSREP: Requesting state transfer: success, donor: 0
2016-07-28 14:55:47 139821049706240 [Note] WSREP: (e31e0085, ‘tcp://0.0.0.0:4567’) turning message relay requesting off
2016-07-28 14:55:48 139821041313536 [Note] WSREP: 0.0 (): State transfer to 1.0 () complete.
2016-07-28 14:55:48 139821041313536 [Note] WSREP: Member 0.0 () synced with group.
WSREP_SST: [INFO] Extracting binlog files: (20160728 14:55:48.132)
mariadb-bin.000038
WSREP_SST: [INFO] Joiner cleanup. rsync PID: 2199 (20160728 14:55:48.137)
WSREP_SST: [INFO] Joiner cleanup done. (20160728 14:55:48.642)
2016-07-28 14:55:48 139821320411264 [Note] WSREP: SST complete, seqno: 5

The State Snapshot Transfer (SST) re-imaged node 2 based on an rsync transfer from node 1, so that explains why node 2 suddenly had a consistent definition of our table.

A DDL Failure in RSU Mode

We’ve seen what happens when DDL fails in TOI mode, but what happens when it fails in RSU mode? This is easy to demonstrate:

MariaDB [db1]> SET wsrep_osu_method='RSU'; Query OK, 0 rows affected (0.00 sec)

MariaDB [db1]> ALTER TABLE tab ADD COLUMN num int DEFAULT NULL;
ERROR 1060 (42S21): Duplicate column name 'num'

In RSU mode, DDL works in similar ways to how it works on a standalone MariaDB/MySQL server, so nothing catastrophic happens when DDL fails. It simply returns an error.

Conclusion

DDL can be kind of weird in Galera Cluster, but many of the quirks are in place to protect the integrity of your data.

Has anyone else noticed strange failures that can happen with DDL in Galera Cluster?

Configuring LDAP Authentication and Group Mapping With MariaDB

Geoff Montee July 14, 2016 MariaDB , MySQL , Security 3 Comments

Author’s note: For the most up-to-date directions on setting up LDAP authentication using PAM and user or group mapping with MariaDB, please see the relevant MariaDB documentation page .

In this blog post, I will demonstrate how to configure MariaDB to use LDAP authentication and group mapping. I have previously written blog posts about configuring PAM authentication and user mapping with MariaDB and configuring PAM authentication and group mapping with MariaDB . If you’ve read those blog posts, a lot of this information will be familiar to you. However, a big difference is that this blog post will also include instructions on setting up an LDAP server.

What do you need to follow these instructions?

A server running MariaDB 10.0+
An RHEL/CentOS 7 server to function as your LDAP server

I am not an LDAP administrator, so if anyone notices that I did anything incorrect or weird, please let me know!

What is group mapping

When I refer to group mapping in this blog post, I am referring to the ability to allow all the members of a POSIX user group to authenticate as a single MariaDB user account. A common use case it to allow all of the members of a DBA-related group to authenticate as a MariaDB superuser account.

The main benefits of this are:

You probably need a POSIX group for your team to control access to shared files anyway, so why not use it to simplify authentication as well?
Even if you have a team of 10 DBAs, you would only need to maintain one MariaDB user account for all of them to share.
There are no shared passwords. Even though there’s only one MariaDB user account, each DBA still uses their own LDAP password to log in.
LDAP is centralized, so even if you have 100 MariaDB servers, group membership only needs to be changed in one place.

In this blog post, I will be mapping the mysql-admins POSIX group to the dba MariaDB user account.

Setting up the LDAP server

If you would like to use LDAP authentication with MariaDB, it is very important that the LDAP Server is set up correctly. The steps in this section have been performed on RHEL 7, but they should be pretty similar for other Linux distributions.

Install LDAP components

First, we need to install the LDAP server and other LDAP components.

sudo yum install openldap openldap-servers openldap-clients nss-pam-ldapd

Create LDAP configuration file from template

Then we need to set up our LDAP configuration. For this, I used a template included with OpenLDAP.

sudo cp /usr/share/openldap-servers/DB_CONFIG.example /var/lib/ldap/DB_CONFIG sudo chown ldap. /var/lib/ldap/DB_CONFIG

Start and enable service

We also want to start the slapd daemon, and make sure that it starts automatically when the system reboots. On RHEL 7, we would execute:

sudo systemctl start slapd sudo systemctl enable slapd

Set the LDAP root password

Then let’s set the root password for the LDAP service. To do that, first we need to use the slappasswd utility to generate a password hash from a clear-text password:

slappasswd

This utility should provide a password hash that looks kind of like this: {SSHA}taDVduzRb34r8wwnhPTDLiYHqwTkHY2k

Now that we have the password hash, let’s create an ldif file to set the root password. LDAP uses ldif files to make changes to the directory.

Let’s make an ldif file to set the LDAP root password using the hash that we created above:

tee ~/olcRootPW.ldif <<EOF dn: olcDatabase={0}config,cn=config changetype: modify add: olcRootPW olcRootPW: {SSHA}taDVduzRb34r8wwnhPTDLiYHqwTkHY2k EOF

Then we can use the ldapadd utility to execute the ldif file:

sudo ldapadd -Y EXTERNAL -H ldapi:/// -f ~/olcRootPW.ldif

Add some standard schemas

OpenLDAP comes with some standard schemas that will be needed later when we want to create POSIX users and groups in our directory. Let’s add those schemas:

sudo ldapadd -Y EXTERNAL -H ldapi:/// -f /etc/openldap/schema/cosine.ldif sudo ldapadd -Y EXTERNAL -H ldapi:/// -f /etc/openldap/schema/nis.ldif sudo ldapadd -Y EXTERNAL -H ldapi:/// -f /etc/openldap/schema/inetorgperson.ldif

Setup the directory manager

Next, let’s set up a directory manager. The directory manager is a privileged LDAP user that we will use to make changes to the directory after this step.

Let’s use the slappasswd utility to generate a password hash from a clear-text password just like we did for the root password above. Simply execute:

slappasswd

Just like it did above, this utility should provide a password hash that looks kind of like this: {SSHA}A0oN2jPVFafjxeb92VwYRwwbZMVppMam

Now that we have the password hash, let’s create an ldif file to create the directory manager:

tee ~/setupDirectoryManager.ldif <<EOF dn: olcDatabase={1}monitor,cn=config changetype: modify replace: olcAccess olcAccess: {0}to * by dn.base="gidNumber=0+uidNumber=0,cn=peercred,cn=external,cn=auth" read by dn.base="cn=Manager,dc=support,dc=mariadb" read by * none

dn: olcDatabase={2}hdb,cn=config
changetype: modify
replace: olcSuffix
olcSuffix: dc=support,dc=mariadb

dn: olcDatabase={2}hdb,cn=config
changetype: modify
replace: olcRootDN
olcRootDN: cn=Manager,dc=support,dc=mariadb

dn: olcDatabase={2}hdb,cn=config
changetype: modify
add: olcRootPW
olcRootPW: {SSHA}A0oN2jPVFafjxeb92VwYRwwbZMVppMam

dn: olcDatabase={2}hdb,cn=config changetype: modify add: olcAccess olcAccess: {0}to attrs=userPassword,shadowLastChange by dn="cn=Manager,dc=support,dc=mariadb" write by anonymous auth by self write by * none olcAccess: {1}to dn.base="" by * read olcAccess: {2}to * by dn="cn=Manager,dc=support,dc=mariadb" write by * read EOF

Note that I am using the dc=support,dc=mariadb domain for my directory. You can change this to whatever is relevant to you.

Now let’s run the ldif file:

sudo ldapmodify -Y EXTERNAL -H ldapi:/// -f ~/setupDirectoryManager.ldif

Setup the base domain

Now let’s create an ldif file to setup the base domain:

tee ~/setupBaseDomain.ldif <<EOF dn: dc=support,dc=mariadb objectClass: top objectClass: dcObject objectclass: organization o: Support Team dc: support

dn: cn=Manager,dc=support,dc=mariadb
objectClass: organizationalRole
cn: Manager
description: Directory Manager

dn: ou=People,dc=support,dc=mariadb
objectClass: organizationalUnit
ou: People

dn: ou=Group,dc=support,dc=mariadb objectClass: organizationalUnit ou: Group EOF

And then run it:

ldapadd -x -D cn=Manager,dc=support,dc=mariadb -W -f ~/setupBaseDomain.ldif

Setup the POSIX group

Above, I mentioned that we would be mapping the mysql-admins POSIX group to the dba MariaDB user. Let’s create an ldif file to represent this group:

tee ~/createMySQLAdminsGroup.ldif <<EOF dn: cn=mysql-admins,ou=Group,dc=support,dc=mariadb objectClass: top objectClass: posixGroup gidNumber: 678 EOF

And then let’s run it:

ldapadd -x -D cn=Manager,dc=support,dc=mariadb -W -f ~/createMySQLAdminsGroup.ldif

Setup a POSIX user

We also need to have a POSIX user account who is a member of our POSIX group. Let’s create an ldif file for a user account named geoff .

tee ~/createGeoffUser.ldif <<EOF dn: uid=geoff,ou=People,dc=support,dc=mariadb objectClass: top objectClass: account objectClass: posixAccount objectClass: shadowAccount cn: geoff uid: geoff uidNumber: 16859 gidNumber: 100 homeDirectory: /home/geoff loginShell: /bin/bash gecos: geoff userPassword: {crypt}x shadowLastChange: -1 shadowMax: -1 shadowWarning: 0 EOF

Then let’s run it:

ldapadd -x -D cn=Manager,dc=support,dc=mariadb -W -f ~/createGeoffUser.ldif

Then set the user’s password:

ldappasswd -x -D cn=Manager,dc=support,dc=mariadb -W -S uid=geoff,ou=People,dc=support,dc=mariadb

Add the user to the group

Both the user and group exist, but the user isn’t yet a member of the group. Let’s create an ldif file to add the user to the group:

tee ~/addMySQLAdminsGroupMembers.ldif <<EOF dn: cn=mysql-admins,ou=Group,dc=support,dc=mariadb changetype: modify add: memberuid memberuid: geoff EOF

And then run it:

ldapmodify -x -D cn=Manager,dc=support,dc=mariadb -W -f ~/addMySQLAdminsGroupMembers.ldif

Setting up the MariaDB server

Now that the LDAP server is configured, we need to setup the MariaDB server. I won’t show how to install MariaDB in this blog post, since there are already many references available for that. Here, I will only show how to get LDAP authentication and group mapping working with an existing MariaDB server.

Install LDAP and PAM libraries

First, we need to make sure that the LDAP and PAM libraries are installed:

sudo yum install openldap-clients nss-pam-ldapd pam pam-devel

Setup authentication

Now that the LDAP client and libraries are installed, we need to update the PAM configuration to use LDAP. We can use the authconfig utility for this. Be sure to replace –ldapserver and –ldapbasedn with values that are relevant for you.

sudo authconfig --enableldap \ --enableldapauth \ --ldapserver=172.31.27.223 \ --ldapbasedn="dc=support,dc=mariadb" \ --enablemkhomedir \ --update

Test new user account

Now that the server is configured to use LDAP authentication, let’s see if our user account works and if the user is a member of the proper groups.

[ec2-user@ip-172-31-22-174 ~]$ su geoff Password: [geoff@ip-172-31-22-174 ec2-user]$ groups users mysql-admins

Looks great so far!

Setup the user mapping plugin

In order to use user or group mapping with MariaDB’s PAM authentication plugin, we need to install an external user mapping plugin for PAM. We can download this plugin from MariaDB’s source code repository, then build it, and then install it:

wget https://raw.githubusercontent.com/MariaDB/server/10.1/plugin/auth_pam/mapper/pam_user_map.c gcc pam_user_map.c -shared -lpam -fPIC -o pam_user_map.so sudo install --mode=0755 pam_user_map.so /lib64/security/

Setup the PAM policy

Let’s create a PAM policy specifically for MariaDB. Since we want to use LDAP and group mapping, we need to make sure that this policy is written to use the PAM modules for LDAP and user mapping plugins. This policy worked for me:

sudo tee /etc/pam.d/mysql <<EOF #%PAM-1.0 auth sufficient pam_ldap.so use_first_pass auth sufficient pam_unix.so nullok try_first_pass auth required pam_user_map.so

account [default=bad success=ok user_unknown=ignore] pam_ldap.so account required pam_unix.so broken_shadow EOF

Configure the user mapping

The user mapping module looks in /etc/security/user_map.conf for its configuration file. Let’s create that file now:

sudo tee /etc/security/user_map.conf <<EOF @mysql-admins: dba EOF

Notice that we use the @ character to prefix group names. If we just wanted to map the geoff user, we could do this instead:

sudo tee /etc/security/user_map.conf <<EOF geoff: dba EOF

Create a local account for the user functioning as the proxy user

Because of the way the PAM authentication plugin for MariaDB works, we need to have a local user account for the MariaDB user functioning as the proxy user. Our proxy user is named dba , so let’s create a local user account with that name.

sudo useradd dba

Allow mysql to read /etc/shadow

Because of the way PAM authentication works, the user running the mysqld process needs to be able to read /etc/shadow . The default user that runs mysqld is usually named mysql . Let’s make sure that this user can read /etc/shadow :

sudo groupadd shadow sudo usermod -a -G shadow mysql sudo chown root:shadow /etc/shadow sudo chmod g+r /etc/shadow

Setup privileges in MariaDB

Now let’s setup our privileges in MariaDB:

-- Install the plugin INSTALL SONAME 'auth_pam';

— Create the “dba” user
CREATE USER ‘dba’@’localhost’ IDENTIFIED BY ‘strongpassword’;
GRANT ALL PRIVILEGES ON *.* TO ‘dba’@’localhost’;

— Create an anonymous catch-all user that will use the PAM plugin and the mysql PAM policy
CREATE USER ”@’localhost’ IDENTIFIED VIA pam USING ‘mysql’;

-- Allow the anonymous user to proxy as the dba user GRANT PROXY ON 'dba'@'localhost' TO ''@'localhost';

Restart MariaDB

Since we changed the group membership of the mysql user, we have to restart mysqld to put the changes into effect:

sudo systemctl restart mariadb

Try it out

Now let’s try it out:

[ec2-user@ip-172-31-22-174 ~]$ mysql -u geoff -h localhost [mariadb] Password: Welcome to the MariaDB monitor. Commands end with ; or \g. Your MariaDB connection id is 7 Server version: 10.1.14-MariaDB MariaDB Server

Type ‘help;’ or ‘\h’ for help. Type ‘\c’ to clear the current input statement.

Since CURRENT_USER() is showing dba@localhost , we know it worked. Awesome!

Conclusion

LDAP authentication and group mapping is very useful for users who want to consolidate account management. However, it can be a little difficult to setup and administrate. I hope this blog post helps simplify it for some!