When the primary node fails, sentinel switching will fail, and the log outputs "next failover delay: I will not start a failover before" How do I fix this problem？
My redis conf ： rename-command FLUSHALL ADMINFLUSHALL rename-command FLUSHDB ADMINFLUSHDB# rename-command CONFIG ADMINCONFIG rename-command KEYS ADMINKEYS
Comment From: Jeremy-Run
Are you sure that the failover was caused by changing the name? The error log you sent is because of multiple failovers within 6 minutes. If you want to know why the failover failed, you have to look at the previous log. And the log you sent is not complete. The complete log will write the time when the next failover can be performed at the end.
Comment From: tanglu86
Are you sure that the failover was caused by changing the name? The error log you sent is because of multiple failovers within 6 minutes. If you want to know why the failover failed, you have to look at the previous log. And the log you sent is not complete. The complete log will write the time when the next failover can be performed at the end.
I am sure that RENAME caused the failover failure。
I set up a new cluster，If the RENAME option modifies the CONFIG command, the failover fails
Comment From: tanglu86
I've tested many times，When RENAME CONFIG is enabled, sentinel failover will exception
Comment From: Jeremy-Run
I'm guessing you also have an error of "Unknown command 'flushall' reading the append only file" in your log file, because the AOF doesn't support "flushall". You can try shutting down AOF and restarting the cluster. This is turned off by changing "appendonly" to "no" in the configuration file. You better know what it means to turn off AOF before you turn it off.
Comment From: tanglu86
I'm guessing you also have an error of "Unknown command 'flushall' reading the append only file" in your log file, because the AOF doesn't support "flushall". You can try shutting down AOF and restarting the cluster. This is turned off by changing "appendonly" to "no" in the configuration file. You better know what it means to turn off AOF before you turn it off.
I don't enabled AOF and RDB ，I have updated the question，RENAME CONFIG ，not RENAME FLUSHALL or RENAME FLUSHDB
Comment From: hwware
@tanglu86 Hi in order to help solve your issue, could you please provide the following detail to us, then we could try to reproduce it and check why your sentinel can not failover, Thanks
- Which version Redis your are using?
- How is your cluster node acritude? such as how many master node, how many replica node, and how many sentinel?
- Could you please detail config file to us?
- As what you say, you do not enable AOF and RDB, do you enable shard or just simple cluster?
Thanks, and looking forward to your feedback.
Comment From: tanglu86
@tanglu86 Hi in order to help solve your issue, could you please provide the following detail to us, then we could try to reproduce it and check why your sentinel can not failover, Thanks
- Which version Redis your are using?
- How is your cluster node acritude? such as how many master node, how many replica node, and how many sentinel?
- Could you please detail config file to us?
- As what you say, you do not enable AOF and RDB, do you enable shard or just simple cluster?
Thanks, and looking forward to your feedback.
- The redis version I use is 5.0.9 2、1 master node, 2 replica nodes, 3 sentinel nodes 3、My profile： redis.conf.txt sentinel.conf.txt
Comment From: hwware
Hi @tanglu86 , I tried replicating this by using some combinations of configs and I was able to get the error message which you received. I will take a deeper dive into this to find the exact cause of this and let you know what I find out.
I used the following configs: master.conf
bind 127.0.0.1 protected-mode yes port 6379 rename-command FLUSHALL ADMINFLUSHALL rename-command FLUSHDB ADMINFLUSHDB rename-command CONFIG ADMINCONFIG rename-command KEYS ADMINKEYS
replica1.conf
bind 127.0.0.1 protected-mode yes port 6380 replicaof 127.0.0.1 6379
rename-command FLUSHALL ADMINFLUSHALL rename-command FLUSHDB ADMINFLUSHDB rename-command CONFIG ADMINCONFIG rename-command KEYS ADMINKEYS
replica2.conf
bind 127.0.0.1 protected-mode yes port 6381 replicaof 127.0.0.1 6379
rename-command FLUSHALL ADMINFLUSHALL rename-command FLUSHDB ADMINFLUSHDB rename-command CONFIG ADMINCONFIG rename-command KEYS ADMINKEYS
sentinel1.conf
port 5000 sentinel monitor mymaster 127.0.0.1 6379 2 sentinel down-after-milliseconds mymaster 6000 sentinel failover-timeout mymaster 18000 replica-announce-ip "127.0.0.1"
sentinel2.conf
port 5001 sentinel monitor mymaster 127.0.0.1 6379 2 sentinel down-after-milliseconds mymaster 6000 sentinel failover-timeout mymaster 18000 replica-announce-ip "127.0.0.1"
sentinel3.conf
port 5002 sentinel monitor mymaster 127.0.0.1 6379 2 sentinel down-after-milliseconds mymaster 6000 sentinel failover-timeout mymaster 18000 replica-announce-ip "127.0.0.1"