Memcached启动脚本

# vim /etc/init.d/memcached
#!/bin/bash
#=======================================================================================
# chkconfig: - 80 12
# description: Distributed memory caching daemon
# processname: memcached
#=======================================================================================
IPADDR=`/sbin/ifconfig eth1 | awk -F ':' '/inet addr/{print $2}' | sed 's/[a-zA-Z ]//g'`
PORT="11211"
USER="root"
SIZE="2048"
CONNNUM="51200"
PIDFILE="/var/run/memcached.pid"
BINFILE="/usr/local/memcached/bin/memcached"
LOCKFILE="/var/lock/subsys/memcached"
RETVAL=0
      
start() {
  echo -n $"Starting memcached......"
  $BINFILE -d -l $IPADDR -p $PORT -u $USER -m $SIZE -c $CONNNUM -P $PIDFILE
  RETVAL=$"Shutting down memcached......"
  /sbin/killproc $BINFILE
  RETVAL=$"Reloading memcached......"
  /sbin/killproc $BINFILE -HUP
  RETVAL=$"$1" in
start)
  start
  ;;
        
stop)
  stop
  ;;
        
restart)
  restart
  ;;
        
condrestart)
  [ -e $LOCKFILE ] && restart
  RETVAL=$"Usage: $0 {start|stop|restart|condrestart|status}"
  RETVAL=1
esac

           

exit $RETVAL
# chmod +x /etc/init.d/memcached
# chkconfig --add memcached
# chkconfig --level 235 memcached on
# service memcached start

binlog 自动清理脚本

# vim /data/scripts/delete_mysql_binlog.sh
#!/bin/bash
#=======================================================================================
#  用于删除 MySQL Master 端已经同步完的 binlog【需在 Master 端运行】,以减少磁盘空间
#  每天凌晨 5:30 分运行一次
#
#  注:需在 Slave 端添加允许 Master 端访问的帐号【帐号:check_binlog,密码:binlog_2356】
#     运行于 MySQL Master 端【目前只用于一主一从的同步模式,对于多从的情况暂时未考虑】
#=======================================================================================
PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin
   
## Slave端连接信息
SLAVE_ADDR="XXX.XXX.XXX.XXX"
SLAVE_USER="check_binlog"
SLAVE_PWD="binlog_2356"
   
LOGFILE="/data/logs/db_sync_info.log"
PINGFILE="/tmp/mysqlping.log"
   
## MySQL状态信息查看命令
SQLCMD="show slave status"
   
#=======================================================================================
## 检查MySQL是否已经运行
if [[ `ps aux | grep mysql[d] | wc -l` -eq 0 ]]; then
  echo The MySQL is not running at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
## 测试Slave端的连通性
nohup mysqladmin -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} ping > ${PINGFILE}
retval=`grep "^error" ${PINGFILE}`
rm -f ${PINGFILE}
if [[ "${retval}X" != "X" ]]; then
  echo The MySQL Slave can not be connected at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
## 检查是否合法的Slave
MASTER_ADDR=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "${SQLCMD}\G;" | awk '$1=="Master_Host:" {print $2}'`
LOCAL_ADDR=`/sbin/ifconfig eth1 | awk -F ':' '/inet addr/{print $2}' | sed 's/[a-zA-Z ]//g'`
if [[ "${MASTER_ADDR}" != "${LOCAL_ADDR}" ]]; then
  echo The MySQL Slave is not lawful at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
## 获得Slave端信息,以此来确定是否处于正常同步的情况
IO_STATUS=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "${SQLCMD}\G;" | awk '$1=="Slave_IO_Running:" {print $2}'`
SQL_STATUS=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "${SQLCMD}\G;" | awk '$1=="Slave_SQL_Running:" {print $2}'`
if [[ "${IO_STATUS}" != "Yes" || "${SQL_STATUS}" != "Yes" ]]; then
  echo The MySQL Replication is not synchronous at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
## 再做一次判断,以保证数据同步绝对正常【创建测试数据】
mysql -uroot -e "create database if not exists mytestdb;"
sleep 3
   
retval=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "show databases;" | grep mytestdb`
mysql -uroot -e "drop database if exists mytestdb;"
if [[ "${retval}X" = "X" ]]; then
  echo The MySQL Replication is not synchronous at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
## 在已经同步的情况,还需要判断当前同步的binlog,以此来确定哪些已经是过期的binlog
SLAVE_BINLOG1=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "${SQLCMD}\G;" | awk '$1=="Master_Log_File:" {print $2}'`
SLAVE_BINLOG2=`mysql -h${SLAVE_ADDR} -u${SLAVE_USER} -p${SLAVE_PWD} -e "${SQLCMD}\G;" | awk '$1=="Relay_Master_Log_File:" {print $2}'`
## 获得Master端,当前的binlog文件以及binlog路径
MASTER_BINLOG=`mysql -uroot -e "show master status;" | grep -v '^+' | tail -1 | awk '{print $1}'`
   
## 主从端已经同步到相同的binlog
if [[ "${SLAVE_BINLOG1}" = "${SLAVE_BINLOG2}" && "${SLAVE_BINLOG1}" = "${MASTER_BINLOG}" ]]; then
  CURR_BINLOG="${MASTER_BINLOG}"
     
## 主从端已经同步,但从端的binlog还没有追赶到主端最新的binlog
elif [[ "${SLAVE_BINLOG1}" = "${SLAVE_BINLOG2}" && "${SLAVE_BINLOG1}" != "${MASTER_BINLOG}" ]]; then
  CURR_BINLOG="${SLAVE_BINLOG1}"
     
## 主从端已经同步,主从端的binlog一致,但relaylog还不一致
elif [[ "${SLAVE_BINLOG1}" != "${SLAVE_BINLOG2}" && "${SLAVE_BINLOG1}" = "${MASTER_BINLOG}" ]]; then
  CURR_BINLOG="${SLAVE_BINLOG2}"
     
else
  echo Has noknown error at:`date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
   
mysql -uroot -e "purge binary logs to '${CURR_BINLOG}';"
if [[ $" "%H-%M-%S`  ${LOGFILE}
fi

# crontab -e 
30 05 * * * /data/scripts/delete_mysql_binlog.sh >/dev/null 2>&1

修复MySQL主从同步

#!/bin/sh
PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin
  
LOGFILE=/data/repair_mysql_sync_`date +%F`.log
SQLCMD1="show slave status"
  
## 查看MySQL是否已启动
if [[ `ps aux | grep mysqld | grep -v grep`"X" = "X" ]]; then
  echo The MySQL is not running at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
  
## 获得MySQL从端Relay binlog的路径
retval=`grep "^relay-log" /etc/my.cnf | grep -v relay-log- | grep '/'`
if [[ "${retval}" = "X" ]]; then
  RELAY_BINLOG_PATH=`ps aux | grep -w mysqld | grep -v grep | awk '{print $13}' | awk -F '=' '{print $2}'`
else
  RELAY_BINLOG_PATH=`dirname $(echo ${retval} | awk -F '=' '{print $2}')`
fi
  
## 查找master.info文件,用于定位Binlog信息
MASTER_FILE=`ps aux | grep -w mysqld | grep -v grep | awk '{print $13}' | awk -F '=' '{print $2}'`/master.info
if [[ ! -e ${MASTER_FILE} ]]; then
  echo This Server is not MySQL Slave at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 1
fi
  
## 获得当前的同步状态
IO_STATUS=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Slave_IO_Running:" {print $2}'`
SQL_STATUS=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Slave_SQL_Running:" {print $2}'`
if [[ "${IO_STATUS}" = "Yes" && "${SQL_STATUS}" = "Yes" ]]; then
  echo Now, The MySQL Replication is synchronous at: `date +%F" "%H-%M-%S`  ${LOGFILE}
  exit 0
fi
  
## 从master.info文件中,获得MySQL主端的同步信息
REPLI_INFO=`sed '/^$/d' ${MASTER_FILE} | tail +2 | head -5`
REPLI_BINLOG_FILE=`echo ${REPLI_INFO} | awk '{print $1}'`
REPLI_IPADDR=`echo ${REPLI_INFO} | awk '{print $3}'`
REPLI_USER=`echo ${REPLI_INFO} | awk '{print $4}'`
REPLI_PWD=`echo ${REPLI_INFO} | awk '{print $5}'`
  
## 删除无用的Relay binlog
rm -rf ${RELAY_BINLOG_PATH}/*-relay-bin.*
  
## 直接从0位置开始同步
SQLCMD2="change master to master_host='${REPLI_IPADDR}', master_user='${REPLI_USER}', master_password='${REPLI_PWD}',"
SQLCMD2="${SQLCMD2} master_log_file='${REPLI_BINLOG_FILE}', master_log_pos=0"
mysql -uroot -e "stop slave;"
mysql -uroot -e "${SQLCMD2};"
mysql -uroot -e "start slave;"
  
## 如果同步的过程中,出现重复记录导致同步失败,就跳过
while true
do
  sleep 2
  IO_STATUS=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Slave_IO_Running:" {print $2}'`
  SQL_STATUS=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Slave_SQL_Running:" {print $2}'`
  BEHIND_STATUS=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Seconds_Behind_Master:" {print $2}'`
  SLAVE_BINLOG1=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Master_Log_File:" {print $2}'`
  SLAVE_BINLOG2=`mysql -uroot -e "${SQLCMD1}\G;" | awk '$1=="Relay_Master_Log_File:" {print $2}'`
    
  ## 出现错误,就将错误信息记录到日志文件,并跳过错误继续同步
  if [[ "${IO_STATUS}" != "Yes" || "${SQL_STATUS}" != "Yes" ]]; then
    ERRORINFO=`mysql -uroot -e "${SQLCMD1}\G;" | awk -F ': ' '$1=="Last_Error" {print $2}'`
    echo "The MySQL synchronous error information: ${ERRORINFO}"  ${LOGFILE}
    mysql -uroot -e "stop slave;"
    mysql -uroot -e "set GLOBAL SQL_SLAVE_SKIP_COUNTER=1;"
    mysql -uroot -e "start slave;"
      
  ## 已完成同步,就正常退出
  elif [[ "${IO_STATUS}" = "Yes" && "${SQL_STATUS}" = "Yes" && "${SLAVE_BINLOG1}" = "${SLAVE_BINLOG2}" && ${BEHIND_STATUS} -eq 0 ]]; then
    echo The MySQL synchronous is ok at: `date +%F" "%H-%M-%S`  ${LOGFILE}
    break
  fi
done

广告合作:本站广告合作请联系QQ:858582 申请时备注:广告合作(否则不回)
免责声明:本站资源来自互联网收集,仅供用于学习和交流,请遵循相关法律法规,本站一切资源不代表本站立场,如有侵权、后门、不妥请联系本站删除!

RTX 5090要首发 性能要翻倍!三星展示GDDR7显存

三星在GTC上展示了专为下一代游戏GPU设计的GDDR7内存。

首次推出的GDDR7内存模块密度为16GB,每个模块容量为2GB。其速度预设为32 Gbps(PAM3),但也可以降至28 Gbps,以提高产量和初始阶段的整体性能和成本效益。

据三星表示,GDDR7内存的能效将提高20%,同时工作电压仅为1.1V,低于标准的1.2V。通过采用更新的封装材料和优化的电路设计,使得在高速运行时的发热量降低,GDDR7的热阻比GDDR6降低了70%。