有一台web服务器异常.
CPU耗尽导致所有请求阻塞,偶发,一般持续时间5-10分钟.

在该时间段,数据库整体负载正常.

使用命令查询慢日志.
指定某个IP,找到查询时间大于100s,并且显示当时执行的SQL

?cat slow-log | sed? '/\#/{N;N;s/\n//g}'? | grep "Query_time: [1-9][0-9]\{2\}\." -n20

或者

cat slow-log | sed? '/\#/{N;N;s/\n//g}'? | grep "127\.0\.0\.1.*Query_time: [1-9][0-9]\{2\}\." -A5

?-bash-4.1$ cat slow-log | sed? '/\#/{N;N;s/\n//g}'? | grep "Query_time: [1-9][0-9]\{2\}\."?
?# Time: 151117 15:49:01# User@Host: xx[xx] @? [IP]? Id: 12827130# Query_time: 103.301519? Lock_time: 0.000080 Rows_sent: 716360? Rows_examined: 1782770
?# Time: 151117 15:57:41# User@Host: xx[xx] @? [IP]? Id: 12827130# Query_time: 102.565315? Lock_time: 0.000088 Rows_sent: 1066410? Rows_examined: 1066410
?# Time: 151117 16:33:51# User@Host: xx[xx] @? [IP]? Id: 12827970# Query_time: 265.482413? Lock_time: 0.000253 Rows_sent: 423799? Rows_examined: 5864759
?# Time: 151117 16:34:02# User@Host: xx[xx] @? [IP]? Id: 12827913# Query_time: 314.793697? Lock_time: 0.000266 Rows_sent: 897670? Rows_examined: 6338629