Giters

sreeix / redis-proxy

A Proxy for redis servers, for failover and standby handling.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Frequently do switching between slave and master

atrun opened this issue · comments

commented

The time interval in about 5 minutes,I worry about performance.

console show debug info:

    debug: Pinging 192.168.80.92:6379
    debug: Pinging 192.168.80.83:6379
    debug: Pinging 192.168.80.83:6380
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    debug: We have a server that went up
    info: Marking 192.168.80.83:6379 as slave of  192.168.80.92: 6379
    debug: OK
    debug: Pinging 192.168.80.92:6379
    debug: Pinging 192.168.80.83:6379
    debug: Pinging 192.168.80.83:6380
    info: end happened undefined
    debug: releaseing[object Object]
    error: Main server down PANIC
    info: finding next active server.
    debug: Creating the pool for active server6379
    info: 192.168.80.83:6379 is slave of no one
    info: Marking 192.168.80.92:6379 as slave of  192.168.80.83: 6379
    info: Marking 192.168.80.83:6380 as slave of  192.168.80.83: 6379
    info: Setting up as active 192.168.80.83 : 6379
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    info: end happened undefined
    debug: releaseing[object Object]
    debug: OK
    debug: OK
    debug: We have a server that went up
    info: Marking 192.168.80.92:6379 as slave of  192.168.80.83: 6379
    debug: OK
    debug: Pinging 192.168.80.92:6379
    debug: Pinging 192.168.80.83:6379
    debug: Pinging 192.168.80.83:6380
commented

You mean that the pinging to the servers is happening more frequently? Or the server is actually going down?

commented

Two kinds of circumstances are not, all servers have not down, but re-selection of the Master, to create connection pool.

[ Creating the pool for active server6379
info: 192.168.80.83:6379 is slave of no one
info: Marking 192.168.80.92:6379 as slave of 192.168.80.83: 6379
info: Marking 192.168.80.83:6380 as slave of 192.168.80.83: 6379]

I analyze the possible connection pool timeout , to create a new connection pool.

commented

Is it still happening on the current version?