12组nodes MySQL DB,每组2台Master-Master,批量清除过期的binl_MySQL
12组nodes MySQL DB,每组2台Master-Master,批量清除过期的binlog释放磁盘空间
接到call,磁盘报警了,发现是大量dml操作导致N多binlo*生,所以需要清除已经sync过的binlog日志,12组nodes MySQL DB,每组2台Master-Master,需要批量清除过期的binlog释放磁盘空间。
方案,总共24台db,一台台进去清理肯定不行,得需要写一个脚本,进行批量操作,方案思路大概如下
1, 建立双master列表masterlist; 一个master一行。
2,远程获取master db上面的binlog位置以及对应master的master主机名(也许是ip地址)
3,拿到binlog位置以及master主机名,然后ssh远程清理掉远程master上面的binlog
4, 采用shell for循环操作step 2以及step 3。
clearbinlog.sh脚本如下
[python]
for masterdb in `cat master.db.full`;do
#1 echo get the binlog position infomation
str_log_files=`ssh $masterdb "/opt/mysql/product/5.5.25a/bin/mysql -uroot --password="" -e /"show slave status/G;/" |grep -i master_Log_File "`
echo $str_log_files;
log_file=`echo $str_log_files | awk '{print $2}'`;
echo $log_file;
#2 echo get the master ip address or master hostname
db01tmp=`ssh $masterdb " /opt/mysql/product/5.5.25a/bin/mysql -uroot --password="" -e /"show slave status/G;/" |grep -i Master_Host "`;
db01=`echo $db01tmp | awk '{print $2}'`
#3 begin to clear the old binlog
ssh $db01 "/opt/mysql/product/5.5.25a/bin/mysql -uroot --password="" -e /"purge master logs to '$log_file';/""
#4 check the disk space for master
ssh $db01 "df -h"
echo " "
echo " -- -- -- ";
done;
OK,run sh脚本
sh clearbinlog.sh 就可以开始清理所有db的binlog了。
最后再次check disk space,执行check_disk.sh脚本,脚本内容如下:
[html]
for masterdb in `master.db.full`;do
ssh $masterdb "df -h" |grep -i mysqldatadir;
done;
执行sh check_disk.sh开始check