驱动故障

基本信息:

  • 业务: 福建农信风控系统

  • 操作系统: 鲲鹏v10 内核4.19.90

  • 资源配置: 8C16G

  • redis版本: redis: 7.0.9

  • redis架构:集群模式 4分片一主一从

  • redis驱动: redisson 3.16.8

  • 节点信息

    172.13.104.1:60145 0- 4095  keys 3211
    172.13.104.18:60145
    
    172.13.104.2:60145 4096-8191 keys 3244
    172.13.104.19:60145
    
    172.13.104.3:60145 8192-12287 keys 3252  *
    172.13.104.20:60145
    
    172.13.104.4:60145 12288-16383 keys 3251
    172.13.104.5:60145
    
  • redis配置

    daemonize yes
    save ""
    port 60145
    cluster-port 60146
    cluster-enabled yes
    cluster-config-file nodes.conf
    cluster-node-timeout 5000
    appendonly yes
    auto-aof-rewrite-percentage 100
    auto-aof-rewrite-min-size 64mb
    no-appendfsync-on-rewrite yes
    pidfile ./redis_60145.pid
    logfile "./redis.log"
    syslog-enabled yes
    #bind 192.13.64.13
    dir ./
    masterauth xxxx
    requlrepass xxxx
    protected-mode no
    ignore-warnings ARM64-COW-BUG
    

故障描述:
应用端的日志提示:org.redisson.client.RedisConnectionException: SlaveConnectionPool no available Redis entries. Master entry host

https://github.com/redisson/redisson/issues/2881

https://github.com/redisson/redisson/issues/3983

问题恢复:

应用服务重启后解决,根因不明。

可能是驱动bug,根据修复日志看该故障有多重原因造成,以下是redission 客户端修复记录:

### 05-Aug-2020 - 3.13.3 released
Fixed - "SlaveConnectionPool no available Redis entries" error is thrown after failover

### 20-Mar-2022 - 3.17.0 released
Fixed - SlaveConnectionPool no available Redis entries error occurs in Cluster mode  

### 24-Aug-2022 - 3.17.6 released
Fixed - "SlaveConnectionPool no available Redis entries" error occurs in Cluster caused by early excluding of master node from nodes for reading 

### 28-Jul-2023 - 3.23.2 released
Fixed - `SlaveConnectionPool no available Redis entries` error may arise in some cases
posted @ 2024-10-14 23:16  mingtian是吧  阅读(27)  评论(0)    收藏  举报