Zabbix分区优化

Posted 2021-05-01 腾讯云TStack

使用zabbix最大的瓶颈在于数据库，维护好zabbix的数据存储，告警，就能很好地应用zabbix去构建监控系统。目前zabbix的数据主要存储在history和trends的2个表中，随着时间的推移，这两个表变得非常大，性能会非常差，影响监控的使用。对mysql进行调优，能够极大的提升Zabbix的性能，本文采用对MySQL进行分区的方法进行调优。

原理

对zabbix中的history和trends等表进行分区，按日期进行分区，每天一个，共保留90天分区。

操作详细步骤

操作影响： 可以在线操作，MySQL的读写变慢，Zabbix性能变慢，影响时间根据数据的小而变化，一般在2个小时左右。

第一步

登录zabbix server的数据库，统一MySQL的配置

cat > /etc/my.cnf<<EOF
[mysqld]
datadir=/data/mysql
socket=/var/lib/mysql/mysql.sock
default-storage-engine = innodb
collation-server = utf8_general_ci
init-connect = 'SET NAMES utf8'
character-set-server = utf8
symbolic-links=0
max_connections=4096
innodb_buffer_pool_size=12G
max_allowed_packet = 32M
join_buffer_size=2M
sort_buffer_size=2M 
query_cache_size = 64M  
query_cache_limit = 4M   
thread_concurrency = 8
table_open_cache=1024
innodb_flush_log_at_trx_commit = 0

long_query_time = 1
log-slow-queries =/data/mysql/mysql-slow.log 

[mysqld_safe]
log-error=/var/log/mariadb/mariadb.log
pid-file=/var/run/mariadb/mariadb.pid

#[mysql]
#socket=/data/mysql/mysql.sock
#
# include all files from the config directory
#
!includedir /etc/my.cnf.d
EOF

注意：一定要修改innodb_buffer_pool_size=物理内存的1/3

第二步

先确认zabbix的版本，本操作zabbix的版本一定要大于3.2.0。小于3.2的版本不能安装此操作，线上默认是zabbix-3.2.6。
 a、 导入存储过程

#cat  partition.sql
DELIMITER $$
CREATE PROCEDURE `partition_create`(SCHEMANAMEvarchar(64), TABLENAME varchar(64), PARTITIONNAME varchar(64), CLOCK int)
BEGIN
       /*
          SCHEMANAME = The DB schema in which to make changes
          TABLENAME = The table with partitions to potentially delete
          PARTITIONNAME = The name of the partition to create
       */
       /*
          Verify that the partition does not already exist
       */

       DECLARE RETROWS INT;
       SELECT COUNT(1) INTO RETROWS
       FROM information_schema.partitions
       WHERE table_schema = SCHEMANAME AND table_name = TABLENAME ANDpartition_description >= CLOCK;

       IF RETROWS = 0 THEN
               /*
                   1. Print a messageindicating that a partition was created.
                   2. Create the SQL to createthe partition.
                   3. Execute the SQL from #2.
               */
               SELECT CONCAT( "partition_create(", SCHEMANAME, ",",TABLENAME, ",", PARTITIONNAME, ",", CLOCK, ")" )AS msg;
               SET @sql = CONCAT( 'ALTER TABLE ', SCHEMANAME, '.', TABLENAME, ' ADDPARTITION (PARTITION ', PARTITIONNAME, ' VALUES LESS THAN (', CLOCK, '));' );
               PREPARE STMT FROM @sql;
               EXECUTE STMT;
               DEALLOCATE PREPARE STMT;
       END IF;
END$$
DELIMITER ;

DELIMITER $$
CREATE PROCEDURE `partition_drop`(SCHEMANAMEVARCHAR(64), TABLENAME VARCHAR(64), DELETE_BELOW_PARTITION_DATE BIGINT)
BEGIN
       /*
           SCHEMANAME = The DB schema in which tomake changes
          TABLENAME = The table with partitions to potentially delete
          DELETE_BELOW_PARTITION_DATE = Delete any partitions with names that aredates older than this one (yyyy-mm-dd)
        */
       DECLARE done INT DEFAULT FALSE;
       DECLARE drop_part_name VARCHAR(16);

       /*
          Get a list of all the partitions that are older than the date
          in DELETE_BELOW_PARTITION_DATE. All partitions are prefixed with
           a "p", so use SUBSTRING TOget rid of that character.
       */
       DECLARE myCursor CURSOR FOR
               SELECT partition_name
               FROM information_schema.partitions
               WHERE table_schema = SCHEMANAME AND table_name = TABLENAME ANDCAST(SUBSTRING(partition_name FROM 2) AS UNSIGNED) <DELETE_BELOW_PARTITION_DATE;
       DECLARE CONTINUE HANDLER FOR NOT FOUND SET done = TRUE;

       /*
          Create the basics for when we need to drop the partition.  Also, create
          @drop_partitions to hold a comma-delimited list of all partitions that
          should be deleted.
       */
       SET @alter_header = CONCAT("ALTER TABLE ", SCHEMANAME,".", TABLENAME, " DROP PARTITION ");
       SET @drop_partitions = "";

       /*
          Start looping through all the partitions that are too old.
       */
       OPEN myCursor;
       read_loop: LOOP
               FETCH myCursor INTO drop_part_name;
               IF done THEN
                        LEAVE read_loop;
               END IF;
               SET @drop_partitions = IF(@drop_partitions = "",drop_part_name, CONCAT(@drop_partitions, ",", drop_part_name));
       END LOOP;
       IF @drop_partitions != "" THEN
                /*
                   1. Build the SQL to drop allthe necessary partitions.
                   2. Run the SQL to drop thepartitions.
                   3. Print out the tablepartitions that were deleted.
               */
               SET @full_sql = CONCAT(@alter_header, @drop_partitions, ";");
               PREPARE STMT FROM @full_sql;
               EXECUTE STMT;
               DEALLOCATE PREPARE STMT;

               SELECT CONCAT(SCHEMANAME, ".", TABLENAME) AS `table`,@drop_partitions AS `partitions_deleted`;
       ELSE
               /*
                   No partitions are beingdeleted, so print out "N/A" (Not applicable) to indicate
                   that no changes were made.
               */
               SELECT CONCAT(SCHEMANAME, ".", TABLENAME) AS `table`,"N/A" AS `partitions_deleted`;
       END IF;
END$$
DELIMITER ;


DELIMITER $$
CREATE PROCEDURE`partition_maintenance`(SCHEMA_NAME VARCHAR(32), TABLE_NAME VARCHAR(32),KEEP_DATA_DAYS INT, HOURLY_INTERVAL INT, CREATE_NEXT_INTERVALS INT)
BEGIN
       DECLARE OLDER_THAN_PARTITION_DATE VARCHAR(16);
       DECLARE PARTITION_NAME VARCHAR(16);
       DECLARE OLD_PARTITION_NAME VARCHAR(16);
       DECLARE LESS_THAN_TIMESTAMP INT;
       DECLARE CUR_TIME INT;

        CALL partition_verify(SCHEMA_NAME,TABLE_NAME, HOURLY_INTERVAL);
       SET CUR_TIME = UNIX_TIMESTAMP(DATE_FORMAT(NOW(), '%Y-%m-%d 00:00:00'));

       SET @__interval = 1;
       create_loop: LOOP
               IF @__interval > CREATE_NEXT_INTERVALS THEN
                        LEAVE create_loop;
               END IF;

               SET LESS_THAN_TIMESTAMP = CUR_TIME + (HOURLY_INTERVAL * @__interval *3600);
               SET PARTITION_NAME = FROM_UNIXTIME(CUR_TIME + HOURLY_INTERVAL *(@__interval - 1) * 3600, 'p%Y%m%d%H00');
               IF(PARTITION_NAME != OLD_PARTITION_NAME) THEN
                        CALLpartition_create(SCHEMA_NAME, TABLE_NAME, PARTITION_NAME, LESS_THAN_TIMESTAMP);
               END IF;
               SET @__interval=@__interval+1;
               SET OLD_PARTITION_NAME = PARTITION_NAME;
       END LOOP;

       SET OLDER_THAN_PARTITION_DATE=DATE_FORMAT(DATE_SUB(NOW(), INTERVALKEEP_DATA_DAYS DAY), '%Y%m%d0000');
       CALL partition_drop(SCHEMA_NAME, TABLE_NAME, OLDER_THAN_PARTITION_DATE);

END$$
DELIMITER ;

DELIMITER $$
CREATE PROCEDURE `partition_verify`(SCHEMANAMEVARCHAR(64), TABLENAME VARCHAR(64), HOURLYINTERVAL INT(11))
BEGIN
       DECLARE PARTITION_NAME VARCHAR(16);
       DECLARE RETROWS INT(11);
       DECLARE FUTURE_TIMESTAMP TIMESTAMP;

       /*
        * Check if any partitions exist for the given SCHEMANAME.TABLENAME.
        */
       SELECT COUNT(1) INTO RETROWS
       FROM information_schema.partitions
       WHERE table_schema = SCHEMANAME AND table_name = TABLENAME ANDpartition_name IS NULL;

       /*
        * If partitions do not exist, go ahead and partition the table
        */
        IFRETROWS = 1 THEN
               /*
                * Take the current date at 00:00:00 and add HOURLYINTERVAL to it.  This is the timestamp below which we willstore values.
                * We begin partitioning based on the beginning of a day.  This is because we don't want to generate arandom partition
                * that won't necessarily fall in line with the desired partition naming(ie: if the hour interval is 24 hours, we could
                * end up creating a partition now named "p201403270600" whenall other partitions will be like "p201403280000").
                */
               SET FUTURE_TIMESTAMP = TIMESTAMPADD(HOUR, HOURLYINTERVAL,CONCAT(CURDATE(), " ", '00:00:00'));
               SET PARTITION_NAME = DATE_FORMAT(CURDATE(), 'p%Y%m%d%H00');

               -- Create the partitioning query
               SET @__PARTITION_SQL = CONCAT("ALTER TABLE ", SCHEMANAME,".", TABLENAME, " PARTITION BY RANGE(`clock`)");
               SET @__PARTITION_SQL = CONCAT(@__PARTITION_SQL, "(PARTITION ",PARTITION_NAME, " VALUES LESS THAN (",UNIX_TIMESTAMP(FUTURE_TIMESTAMP), "));");

               -- Run the partitioning query
               PREPARE STMT FROM @__PARTITION_SQL;
               EXECUTE STMT;
               DEALLOCATE PREPARE STMT;
       END IF;
END$$
DELIMITER ;

DELIMITER $$
CREATE PROCEDURE`partition_maintenance_all`(SCHEMA_NAME VARCHAR(32))
BEGIN
               CALL partition_maintenance(SCHEMA_NAME, 'history', 90, 24, 14);
               CALL partition_maintenance(SCHEMA_NAME, 'history_log', 90, 24, 14);
               CALL partition_maintenance(SCHEMA_NAME, 'history_str', 90, 24, 14);
               CALL partition_maintenance(SCHEMA_NAME, 'history_text', 90, 24, 14);
                CALLpartition_maintenance(SCHEMA_NAME, 'history_uint', 90, 24, 14);
               CALL partition_maintenance(SCHEMA_NAME, 'trends', 730, 24, 14);
               CALL partition_maintenance(SCHEMA_NAME, 'trends_uint', 730, 24, 14);
END$$
DELIMITER ;

上面内容包含了创建分区的存储过程，将上面内容复制到partition.sql中,然后执行如下：

mysql  -uzabbix -pzabbix zabbix  < partition.sql

b、 添加crontable,每天执行01点01分执行，如下：

crontab  -l > crontab.txt
cat >> crontab.txt <<EOF
#zabbix partition_maintenance
01 01 * * *  mysql  -uzabbix -pzabbix zabbix -e"CALL partition_maintenance_all('zabbix')" &>/dev/null
EOF
cat crontab.txt |crontab

注意： mysql的zabbix用户的密码部分按照实际环境配置

c、首先执行一次（由于首次执行的时间较长，请使用nohup执行），如下：

nohup   mysql -uzabbix -pzabbix zabbix -e "CALLpartition_maintenance_all('zabbix')" &> /root/partition.log&  

注意：观察/root/partition.log的输出

d、 查看结果
 登录mysql,查看history等表, 如下：

MariaDB [zabbix]> showcreate table history
| history | CREATE TABLE `history` (
 `itemid` bigint(20) unsigned NOT NULL,
  `clock`int(11) NOT NULL DEFAULT '0',
  `value`double(16,4) NOT NULL DEFAULT '0.0000',
  `ns`int(11) NOT NULL DEFAULT '0',
  KEY`history_1` (`itemid`,`clock`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8
/*!50100 PARTITION BY RANGE (`clock`)
(PARTITION p201708280000 VALUES LESS THAN(1503936000) ENGINE = InnoDB,
 PARTITION p201708290000 VALUES LESS THAN(1504022400) ENGINE = InnoDB,
 PARTITION p201708300000 VALUES LESS THAN(1504108800) ENGINE = InnoDB,
 PARTITION p201708310000 VALUES LESS THAN(1504195200) ENGINE = InnoDB,
 PARTITION p201709010000 VALUES LESS THAN(1504281600) ENGINE = InnoDB,
 PARTITION p201709020000 VALUES LESS THAN(1504368000) ENGINE = InnoDB,
 PARTITION p201709030000 VALUES LESS THAN(1504454400) ENGINE = InnoDB,
 PARTITION p201709040000 VALUES LESS THAN(1504540800) ENGINE = InnoDB,
 PARTITION p201709050000 VALUES LESS THAN(1504627200) ENGINE = InnoDB,
 PARTITION p201709060000 VALUES LESS THAN(1504713600) ENGINE = InnoDB,
 PARTITION p201709070000 VALUES LESS THAN(1504800000) ENGINE = InnoDB,
 PARTITION p201709080000 VALUES LESS THAN(1504886400) ENGINE = InnoDB,
 PARTITION p201709090000 VALUES LESS THAN(1504972800) ENGINE = InnoDB,
 PARTITION p201709100000 VALUES LESS THAN(1505059200) ENGINE = InnoDB,
 PARTITION p201709110000 VALUES LESS THAN(1505145600) ENGINE = InnoDB) */ |

发现了大量PARTITION字段，说明配置正确。注意观察Mysql的Slow Query,一般到执行操作的第二天，Slow Query几乎就会有了，此时Zabbix的Dashboard响应速度应该非常流畅了。