Redis proxy has a lot of CLOSE_WAIT

[2hang5ong]

After the master node in the redis cluster exits abnormally, a large number of "CLOSE_WAIT" connections cannot be disconnected in the pod running redis proxy. This situation will continue until the redis proxy pod is restarted. During this period, the reids client cannot normally connect to the redis cluster through redis proxy. Has anyone encountered the same situation?

(PS: redis proxy and redis clusters， running in k8s cluster)
this my configuration file：
“
cluster redis-service:6379 # 配置为Redis Cluster Service
bind 0.0.0.0
port 12306
threads 8
daemonize no
enable-cross-slot yes
#auth P@ssw0rd
log-level error
”

中文描述：
当redis集群中的master节点异常退出之后，在运行有redis-proxy的pod中，会存在大量的 “CLOSE_WAIT”状态的连接无法被断开，这种情况会一直持续到redis-proxy这个pod重启，在此期间，reids客户端无法正常通过redis-proxy 连接到redis集群,请问有人遇到过相同的情况么？
（ps: redis-proxy以及redis集群，运行在k8s集群中）
这是我的配置文件示例：
"
cluster redis-service:6379 # 配置为Redis Cluster Service
bind 0.0.0.0
port 12306 # redis-cluster-proxy 对外暴露端口
threads 8 # 线程数量
daemonize no
enable-cross-slot yes
#auth P@ssw0rd # 配置Redis Cluster 认证密码
log-level error

"

[huxiaozhe]

可能是私有连接池的问题，然后你后端配了 timeout，不过不建议用这个proxy，bug太多了。