Redis 集群模式优化和备份策略

 vm.swappiness=10

vm.dirty_background_ratio = 5
vm.dirty_ratio = 20

fs.file-max = 2000000

net.core.somaxconn = 1024

net.ipv4.tcp_fin_timeout = 30

net.ipv4.tcp_tw_reuse = 1

net.core.rmem_default = 262144
net.core.wmem_default = 262144
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 65536 16777216

# 安装 cpupower（如果还没装）
dnf install kernel-tools
# 设置 governor
cpupower frequency-set --governor performance

echo none > /sys/block/nvme0n1/queue/scheduler

# /etc/sysctl.d/99-performance.conf
# === 内存管理 ===
vm.swappiness = 10
vm.dirty_ratio = 20
vm.dirty_background_ratio = 5
# 可能还加 dirty_expire / writeback，根据具体测试调优
# vm.dirty_expire_centisecs = 500
# vm.dirty_writeback_centisecs = 100
# 可选：如果你知道内存不是特别紧张，可设置 overcommit
# vm.overcommit_memory = 1
# === 文件句柄限制 ===
fs.file-max = 2000000
# === 网络 / tcp 参数 ===
net.core.somaxconn = 8192  # 监听 backlog
net.core.netdev_max_backlog = 50000
# 接收 / 发送缓冲
net.core.rmem_default = 262144
net.core.wmem_default = 262144
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096   87380   16777216
net.ipv4.tcp_wmem = 4096   65536   16777216
# time_wait 和连接重用
net.ipv4.tcp_fin_timeout = 30
net.ipv4.tcp_tw_reuse = 1
# keepalive
net.ipv4.tcp_keepalive_time = 600
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_keepalive_probes = 3
# 网络安全相关
kernel.sysrq = 0
net.ipv4.ip_forward = 0
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
net.ipv4.conf.all.accept_source_route  =0
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.conf.all.accept_redirects = 0
net.ipv4.conf.default.accept_redirects = 0
net.ipv4.conf.all.secure_redirects = 0
net.ipv4.conf.default.secure_redirects = 0
net.ipv4.icmp_echo_ignore_broadcasts = 1
net.ipv4.icmp_ignore_bogus_error_responses = 1
net.ipv4.conf.all.rp_filter = 1
net.ipv4.conf.default.rp_filter = 1
net.ipv4.tcp_syncookies = 1
kernel.dmesg_restrict = 1
net.ipv6.conf.all.accept_redirects = 0
net.ipv6.conf.default.accept_redirects = 0
kernel.threads-max = 655350
# === 其他网络 /系统选项 ===
# 可以根据需要打开更多，但这里是基础

sysctl --system

########################################
# redis 性能 +集群 配置示例
# 这个配置文件只做参考 不能直接拿来用 redis 是不允许再参数后买你添加注释的直接使用会报错
########################################
# 内存管理
maxmemory 12gb           # 假设节点内存 16g，为 redis 留约 12g
maxmemory-policy allkeys-lru   # 使用 lru 淘汰所有键（适合缓存 + 数据混合场景）
maxmemory-samples 5
# 延迟释放
lazyfree-lazy-eviction yes
lazyfree-lazy-expire yes
lazyfree-lazy-server-del yes
# 持久化 (rdb + aof)
appendonly yes
appendfilename "appendonly.aof"
appendfsync everysec
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
aof-load-truncated yes
# rdb 快照 (可根据需要)
save 900 1
save 300 10
save 60 10000
# 集群配置 (如果是 redis cluster)
cluster-enabled yes
cluster-config-file nodes.conf
cluster-node-timeout 15000
cluster-replica-validity-factor 10
cluster-migration-barrier 1
cluster-require-full-coverage no
# 性能 /调度相关
hz 10
dynamic-hz yes
# lua 脚本执行限制
lua-time-limit 5000  # 毫秒
# 慢日志 (监控慢命令)
slowlog-log-slower-than 10000   # 微秒 (10 ms) — 根据业务调整
slowlog-max-len 128
# 内存碎片 / 编码优化
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-entries 512
list-max-ziplist-value 64
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
# 安全性 (可选，根据场景启用)
# requirepass yourpassword
# rename-command flushall ""
# 其他 (可选，根据需求)
notify-keyspace-events ""

/usr/local/redis/bin/redis-cli -h 192.168.1.11 -p 6379 -a ugwfssnuzgs4e62e cluster nodes|grep master | grep -v "fail" | awk '{print $2}' | cut -d@ -f1

#!/bin/bash
redis_ip="192.168.1.11"         # 任意一个集群节点（用于获取节点列表）
redis_port="6379"
redis_pwd="ugwfssnuzgs4e62e"
backup_dir="/root/backup/"
timeout=300  
mkdir -p "$backup_dir" || { echo "error: 无法创建备份目录 $backup_dir"; exit 1; }
# 获取所有 master 节点
nodes=$(redis-cli -c -h "$redis_ip" -p "$redis_port" -a "$redis_pwd" cluster nodes | \
  grep master | grep -v fail | awk '{print $2}' | cut -d@ -f1)
for node in $nodes; do
  ip=$(echo "$node" | cut -d: -f1)
  port=$(echo "$node" | cut -d: -f2)
  echo "=== 开始备份：redis 节点 $ip:$port ==="
  # 触发 bgsave (rdb)
  redis-cli -h "$ip" -p "$port" -a "$redis_pwd" bgsave
  if [ $? -ne 0 ]; then
    echo "warning: bgsave 在 $ip:$port 执行失败"
  fi
  # 等待 bgsave 完成
  start_time=$(date +%s)
  while true; do
    inprog=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" info persistence | \
      grep rdb_bgsave_in_progress | cut -d: -f2 | tr -d '\r')
    if [ "$inprog" = "0" ]; then
      echo "bgsave 完成：$ip:$port"
      break
    fi
    now=$(date +%s)
    elapsed=$(( now - start_time ))
    if [ $elapsed -gt $timeout ]; then
      echo "error: bgsave 超时 ($elapsed 秒) 在 $ip:$port" >&2
      break
    fi
    echo "等待 bgsave 完成 ($elapsed 秒) ..."
    sleep 5
  done
  # 获取 rdb 文件路径
  dir=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" config get dir | awk 'nr==2')
  dbfile=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" config get dbfilename | awk 'nr==2')
  src_rdb="${dir}/${dbfile}"
  if [ ! -f "$src_rdb" ]; then
    echo "error: 未找到 rdb 文件: $src_rdb" >&2
  else
    dst_rdb="${backup_dir}/${ip}_${port}_$(date +%y%m%d_%h%m%s).rdb"
    cp "$src_rdb" "$dst_rdb" && echo "rdb 备份保存到 $dst_rdb"
  fi
  aof_enabled=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" config get appendonly | awk 'nr==2')
  if [ "$aof_enabled" = "yes" ]; then
    aofdir=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" config get appenddirname | awk 'nr==2')
    aofname=$(redis-cli -h "$ip" -p "$port" -a "$redis_pwd" config get appendfilename | awk 'nr==2')
    if [ -z "$aofdir" ]; then
      src_aof="${dir}/${aofname}"
      if [ -f "$src_aof" ]; then
        dst_aof="${backup_dir}/${ip}_${port}_$(date +%y%m%d_%h%m%s).aof"
        cp "$src_aof" "$dst_aof" && echo "aof 备份保存到 $dst_aof"
      else
        echo "error: aof 文件未找到: $src_aof" >&2
      fi
    else
      # multi-part aof 模式 (redis 7+)
      aof_dir_full="${dir}/${aofdir}"
      if [ -d "$aof_dir_full" ]; then
        dst_mp_dir="${backup_dir}/${ip}_${port}_aof_$(date +%y%m%d_%h%m%s)"
        mkdir -p "$dst_mp_dir"
        cp "${aof_dir_full}"/* "$dst_mp_dir"/ && echo "mp-aof 备份到 $dst_mp_dir"
      else
        echo "error: aof 目录不存在: $aof_dir_full" >&2
      fi
    fi
  else
    echo "aof 未启用在 $ip:$port，跳过 aof 备份"
  fi
done
echo "=== 所有节点备份完成 ==="

# 或者停止对应的 redis 服务 / systemd 单元
redis-cli -h <ip> -p <port> shutdown

cp /backup/redis/cluster/<ip>_<port>_yyyymmdd_hhmmss.rdb /var/lib/redis/$port/dump.rdb
# 如果是 aof
cp /backup/redis/cluster/<ip>_<port>_yyyymmdd_hhmmss.aof /var/lib/redis/$port/appendonly.aof

chown redis:redis /var/lib/redis/$port/dump.rdb
chown redis:redis /var/lib/redis/$port/appendonly.aof

systemctl start redis

redis-cli -h <ip> -p <port> bgrewriteaof

redis-cli -h <恢复节点 ip> -p <恢复节点 端口> cluster meet <任意健康节点 ip> <端口>

redis-cli -h <恢复节点 ip> -p <恢复端口> cluster nodes

redis-cli --cluster reshard <一个集群节点 ip:端口> --from <其他节点列表，用逗号分隔> --to <恢复节点 id> --slots <slot 数量> --yes