how to optimize mysql for high traffic drupal site

本文探讨了如何通过调整MySQL配置来优化Drupal站点性能,包括改变表引擎为InnoDB以减少锁定问题,发现并优化慢查询,合理设置my.cnf参数如连接数限制、等待超时等,以及调整查询缓存大小等措施。适用于处理高流量和数据库密集型网站。

Configuring Mysql could be a daunting task if your Drupal site begins to scale and you start getting "too many connections" error or mysql server down errors. We had this opportunity to scale a Drupal site to 2 million page views a month and we improved this with success. So, lets start and share what we found out. I would be glad to hear from other people's experiences as well in the comment section.

 

Drupal 6 comes with MyISAM engine default. First thing you would notice with your drupal site would be your tables locking. Site would begin to go down and when you look at the mysql processes (using mysql tools or command line), I know this would feel like cursing Drupal or Mysql, but before you begin to hit at Drupal and feeling bad about Drupal performance, please try and change your Drupal tables to InnoDB

 

You would not be able to change all the tables like node revision if similar module is used. Node revision is dependent on MySQL's FULLTEXT indexing for MyISAM tables if similar module is enabled. FULLTEXT does not work with InnoDB.

 

    ALTER  TABLE  `node`   ENGINE  = InnoDB

Locking would go away. MYISAM locks table whenever an update/insert/delete operation is happening on a node table for example. This keeps other operations on wait. When we have too many such operations in queue, this locking builds up and brings the site to halt. InnoDB does locking at the record level, thus preventing the above case.

 

Discover Mysql Slow Queries:

 

This should be the next step to improve mysql performance. Find out the sql queries that are slow or repeating. This is done as follows:

 

a. Use Devel module. This would list the sql queries that are being executed when a page loads. Monitor the unusal slow sql query or queries repeating. Ask yourself: Is there an sql query that can be optimized or can be removed?

 

b. Log Mysql slow queries as follows in /etc/my.cnf

 
log-slow-queries=/var/lib/mysqllogs/slow-log
long_query_time=2
log-queries-not-using-indexes

 

Configuring my.cnf:

Here comes the complex part. This may be tricky. 

 

Here is my my.cnf:

 

a.  Limit mysql connections. This is important to set. If it is too high, your server would stop responding and you would have to reboot.  If it too less, you would start getting "too many connections" error on your Drupal site.  I keep it high enough that I never get too many connections error.

 

max_connections=100

 

 

b. If you see too many sleep connections in your mysql processes (SHOW PROCESSLIST), this means that the there is no mysql operation but the mysql connection is still retained. This prevents the mysql connection to be released and taken up by some other process. You would need to reduce the wait_timeout.

wait_timeout=600

 

c. If you changed to InnoDB, make the following changes to my.cnf:

[mysqld]
 # InnoDB settings
 innodb_data_home_dir = /var/lib/mysql/
 innodb_data_file_path = ibdata1:100M:autoextend
 set-variable = innodb_buffer_pool_size=100M
 set-variable = innodb_additional_mem_pool_size=10M
 innodb_flush_log_at_trx_commit=1 

 

The innodb_data_home_dir setting defines the location where InnoDB should create data files. The innodb_data_file_path setting defines the name of the data files. In this case, it will create a 100MB data file called ibdata1 and will extend the size as needed. A data file in InnoDB parlance is a tablespace.

 

d.Logging slow sql queries:

 

log-slow-queries=/var/lib/mysqllogs/slow-log
long_query_time=2
log-queries-not-using-indexes
 

e. Max allowed packet to be increased from 32M to 128M. Typically, if you are not storing your cache in memcache and in mysql tables, you would need to increase this size.

 

 

max_allowed_packet=128M

 

f. Query Cache size to be adjusted:

query_cache_size=4M
query_cache_limit = 2M
sort_buffer_size=2M
read_buffer_size=2M
read_rnd_buffer_size=4M

 

For a high traffic site which is database intensive and has a dedicated 16GB server to mysql, my my.cnf looks like:

 

[mysqld]
datadir=/var/lib/mysql
tmpdir=/var/lib/mysqltmp
socket=/var/lib/mysql/mysql.sock
skip-locking
skip-name-resolve
table_cache=2048
thread_cache_size=32
back_log=100
max_connect_errors=10000
open-files=10000
interactive_timeout=400
wait_timeout=300
max_connections=500
skip-bdb
log-slow-queries=/var/lib/mysqllogs/slow-log
long_query_time=2
log-queries-not-using-indexes
max_allowed_packet=128M
tmp_table_size=256M
max_heap_table_size=256M
query_cache_size=32M
query_cache_limit = 6M
sort_buffer_size=4M
read_buffer_size=4M
read_rnd_buffer_size=16M
join_buffer_size=2M
default-storage-engine=InnoDB
key_buffer_size=128M
myisam_sort_buffer_size=64M
innodb_log_file_size=100M
innodb_buffer_pool_size=6G
innodb_additional_mem_pool_size=20M
innodb_support_xa = 0
[mysql.server]
user=mysql
[mysqld_safe]
根据原作 https://pan.quark.cn/s/459657bcfd45 的源码改编 Classic-ML-Methods-Algo 引言 建立这个项目,是为了梳理和总结传统机器学习(Machine Learning)方法(methods)或者算法(algo),和各位同仁相互学习交流. 现在的深度学习本质上来自于传统的神经网络模型,很大程度上是传统机器学习的延续,同时也在不少时候需要结合传统方法来实现. 任何机器学习方法基本的流程结构都是通用的;使用的评价方法也基本通用;使用的一些数学知识也是通用的. 本文在梳理传统机器学习方法算法的同时也会顺便补充这些流程,数学上的知识以供参考. 机器学习 机器学习是人工智能(Artificial Intelligence)的一个分支,也是实现人工智能最重要的手段.区别于传统的基于规则(rule-based)的算法,机器学习可以从数据中获取知识,从而实现规定的任务[Ian Goodfellow and Yoshua Bengio and Aaron Courville的Deep Learning].这些知识可以分为四种: 总结(summarization) 预测(prediction) 估计(estimation) 假想验证(hypothesis testing) 机器学习主要关心的是预测[Varian在Big Data : New Tricks for Econometrics],预测的可以是连续性的输出变量,分类,聚类或者物品之间的有趣关联. 机器学习分类 根据数据配置(setting,是否有标签,可以是连续的也可以是离散的)和任务目标,我们可以将机器学习方法分为四种: 无监督(unsupervised) 训练数据没有给定...
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
红包 添加红包
表情包 插入表情
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值