MySQL调优脚本tuning-primer.sh

MySQL调优脚本tuning-primer.sh官方下载地址：http://forge.mysql.com/projects/project.php?id=44 ->https://launchpad.net/mysql-tuning-primer, 直接下载地址： http://launchpad.net/mysql-tuning-primer/trunk/1.5-r5/+download/tuning-primer.sh

该脚本使用 “SHOW STATUS LIKE…” 和 “SHOW VARIABLES LIKE…” 命令获得MySQL相关变量和运行状态。然后根据推荐的调优参数对当前的MySQL数据库进行测试。最后根据不同颜色的标识来提醒用户需要注意的各个参数 设置。该版本兼容 MySQL 3.23 和更高版本（包含 5.1）

当前版本会处理如下这些推荐的参数：

[Copy to clipboard]View Code

1 2 3 4 5 6 7 8 9 10 11 12 

 Slow Query Log （慢查询日志）  Max Connections （最大连接数）  Worker Threads （工作线程）  Key Buffer （Key 缓冲）  Query Cache （查询缓存）  Sort Buffer （排序缓存）  Joins （连接）  Temp Tables （临时表）  Table (Open & Definition) Cache （表缓存）  Table Locking （表锁定）  Table Scans (read_buffer) （表扫描，读缓冲）  Innodb Status （Innodb 状态）

tuning-primer.sh 有如下运行参数:

[Copy to clipboard]View Code

1 2 3 4 5 6 7 

Usage: ./tuning-primer.sh [ mode ] all            运行所有检测(默认值) prompt         显示提示信息 mem, memory    运行有关内存使用方面的检测 disk, file     运行有关I/O性能和文件处理限制方面的检测 innodb         运行InnoDB检测 misc           其它

脚本支持my.cnf文件，可以把user, password, host, socket等连接参数放在~/.my.cnf文件里，如果使用了自定义的socket文件，请修改tuning-primer.sh文件中关于socket文件位置的变量。

运行 tuning-primer.sh显示说明

SLOW QUERIES慢查询检查SLOW QUERIESThe slow query log is enabled.说明我已经启用了慢查询记录功能。也就是参数
slow_query_log = 1Current long_query_time = 5.000000 sec.慢查询的阀值时间。也就是参数
long_query_time = 5You have 17 out of 638844 that take longer than 5.000000 sec. to complete说明慢查询日志中记录了17条查询时间超过5秒的语句。
slow_query_log_file=/data/ats_db/mysql-slow.log设置慢查询日志路径。使用

mysqldumpslow命令查询慢日志

Your long_query_time seems to be fine慢查询阀值时间设置得在推荐的范围内 BINARY UPDATE LOG更新二进制日志文件The binary update log is enabled这项说明启用了bin-log日志功能。参数
log-bin = /data/ats_db/mysql-binBinlog sync is not enabled, you could loose binlog records during a server crash没有启用 sync_binlog 选项。也即是将二进制日志实时写入到磁盘通过 sync_binlog=1来指定 WORKER THREADS工作线程Current thread_cache_size = 8当前线程缓存大小。
thread_concurrency = 8Current threads_cached = 7Show status like ‘threads_cached’Current threads_per_sec = 0脚本先执行Show status like ‘Threads_cached’查看当前的线程创建情况，然后sleep 1后在执行相同的命令，最终后者减去前者的数就是每秒线程创建数。Historic threads_per_sec = 0该值是使用Threads_cached /uptime获得的。Your thread_cache_size is fine  MAX CONNECTIONS最大连接数Current max_connections = 1024当前配置文件中设置的并发连接数Current threads_connected = 2当前线程连接诶数。
show status like ‘Threads_connected’Historic max_used_connections = 4show status like ‘Max_used_connections’;The number of used connections is 0% of the configured maximum.这个值使用 Max_used_connections*100/ max_connections得出。You are using less than 10% of your configured max_connections. Lowering max_connections could help to avoid an over-allocation of memory

See “MEMORY USAGE” section to make sure you are not over-allocating

Max_used_connections的值不足max_connections值的10%。设置合适的max_connections值有助于节省内存。 MEMORY USAGE内存使用Max Memory Ever Allocated : 841 MMax Memory Ever Allocated = max_memoryConfigured Max Per-thread Buffers : 28.40 GConfigured Max Per-thread Buffers = per_thread_buffersConfigured Max Global Buffers : 586 MConfigured Max Global Buffers = per_thread_max_buffersConfigured Max Memory Limit : 28.97 GConfigured Max Memory Limit = total_memory
这 一项很重要，他是将各个缓存的大小累加，然后同max_connections相乘，从而得出当达到max_connections后需要分配的内存有多 少。我这里由于max_connections写得很大，造成了最大内存限制超过了真实内存很多，所以建议不要随意增大max_connections的 值。减小 max_connections的值，最终保证最大内存限制在真实内存的90%以下。Physical Memory : 7.79 G实际物理内存Max memory limit exceeds 90% of physical memory per_thread_buffers
(read_buffer_size+read_rnd_buffer_size +sort_buffer_size+thread_stack+
join_buffer_size+binlog_cache_size)*max_connectionsper_thread_max_buffers
(read_buffer_size+read_rnd_buffer_size +sort_buffer_size+thread_stack
+join_buffer_size+binlog_cache_size)*max_used_connections

global_buffers
innodb_buffer_pool_size+innodb_additional_mem_pool_size+innodb_log_buffer_size+
key_buffer_size+query_cache_size

max_memory=global_buffers+per_thread_max_buffers

total_memory=global_buffers+per_thread_buffers

 KEY BUFFERKey 缓冲Current MyISAM index space = 222 K当前数据库MyISAM表中索引占用磁盘空间Current key_buffer_size = 512 MMySQL配置文件中key_buffer_size 设置的大小Key cache miss rate is 1 : 3316Key_read_requests/ Key_reads 这里说明3316次读取请求中有1次丢失(也就是说1次读取磁盘)Key buffer free ratio = 81 %key_blocks_unused * key_cache_block_size / key_buffer_size * 100Your key_buffer_size seems to be fine  QUERY CACHEQuery 缓存Query cache is enabled该项说明 我们指定了query_cache_size 的值。如果query_cache_size=0的话这里给出的提示是：
Query cache is supported but not enabled
Perhaps you should set the query_cache_sizeCurrent query_cache_size = 64 M当前系统query_cache_size 值大小 [F]Current query_cache_used = 1 Mquery_cache_used =query_cache_size-qcache_free_memoryCurrent query_cache_limit = 128 M变量 query_cache_limit 大小Current Query cache Memory fill ratio = 1.79 %query_cache_used/query_cache_size *100%Current query_cache_min_res_unit = 4 Kshow variables like ‘query_cache_min_res_unit’;Your query_cache_size seems to be too high.
Perhaps you can use these resources elsewhere这项给出的结论是query_cache_size的值设置的有些过高。其比对标准是 “Query cache Memory fill ratio”的值如果小于<25%就会给出这个提示。可以将这些资源应用到其他的地方MySQL won’t cache query results that are larger than query_cache_limit in sizeMySQL不会将大于query_cache_limit的查询结果进行缓存show status like ‘Qcache%’;Qcache_free_blocks        10
Qcache_free_memory        65891984
Qcache_hits            14437
Qcache_inserts            707
Qcache_lowmem_prunes    0
Qcache_not_cached        216
Qcache_queries_in_cache    540
Qcache_total_blocks        1191 SORT OPERATIONSSORT 选项Current sort_buffer_size = 6 Mshow variables like ’sort_buffer%’;Current read_rnd_buffer_size = 16 Mshow variables like ‘read_rnd_buffer_size%’;Sort buffer seems to be fine  JOINSJOINSCurrent join_buffer_size = 132.00 Kshow variables like ‘join_buffer_size%’;join_buffer_size=join_buffer_size+4kbYou have had 6 queries where a join could not use an index properly这里的6是通过 show status like ‘Select_full_join’; 获得的You should enable “log-queries-not-using-indexes”
Then look for non indexed joins in the slow query log.
If you are unable to optimize your queries you may want to increase your
join_buffer_size to accommodate larger joins in one pass.Note! This script will still suggest raising the join_buffer_size when
ANY joins not using indexes are found.你需要启用 “log-queries-not-using-indexes” 然后在慢查询日志中看是否有取消索引的joins语句。如果不优化查询语句的话，则需要增大join_buffer_size。 OPEN FILES LIMIT文件打开数限制Current open_files_limit = 1234 filesshow variables like ‘open_files_limit%’;The open_files_limit should typically be set to at least 2x-3xthat of table_cache if you have heavy MyISAM usage.如果系统中有很多的MyISAM类型的表，则建议将open_files_limit 设置为2X~3X的table_open_cache
show status like ‘Open_files’;open_files_ratio= open_files*100/open_files_limit
如果open_files_ratio 超过75% 则需要加大open_files_limitYour open_files_limit value seems to be fine  TABLE CACHETABLE 缓存Current table_open_cache = 512 tablesshow variables like ‘table_open_cache’;Current table_definition_cache = 256 tablesshow variables like ‘ table_definition_cache ‘;You have a total of 368 tablesSELECTCOUNT(*) FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE=’BASE TABLE’You have 371 open tables.show status like ‘Open_tables’;The table_cache value seems to be fineOpen_tables /table_open_cache*100% < 95%You should probably increase your table_definition_cache value.table_cache_hit_rate =open_tables*100/opened_tables TEMP TABLES临时表Current max_heap_table_size = 16 Mshow variables like ‘max_heap_table_size’;Current tmp_table_size = 16 Mshow variables like ‘tmp_table_size’;Of 285 temp tables, 11% were created on diskCreated_tmp_tables=285created_tmp_disk_tables*100/
(created_tmp_tables+created_tmp_disk_tables)=11%Created disk tmp tables ratio seems fine  TABLE SCANS扫描表Current read_buffer_size = 6 Mshow variables like ‘read_buffer_size’;Current table scan ratio = 9 : 1read_rnd_next =show global status like ‘Handler_read_rnd_next’;
com_select= show global status like ‘Com_select’;
full_table_scans=read_rnd_next/com_select
Current table scan ratio = full_table_scans : 1″
如果表扫描率超过4000，说明进行了太多表扫描，很有可能索引没有建好，增加read_buffer_size值会有一些好处，但最好不要超过8MB。read_buffer_size seems to be fine  TABLE LOCKINGTABLE LOCKINGCurrent Lock Wait ratio = 0 : 5617show global status like’Table_locks_waited’;
show global status like

‘Questions’;
如果 Table_locks_waited=0
Current Lock Wait ratio = 0: Questions

Your table locking seems to be fine