[redis sentinel] failed listening on port 26379 (TCP) & sentinel mode no response problem solved

problem 1 Failed listening on port 26379 (TCP)

The error information is as follows , This problem is usually caused by the port being occupied , Kill the sentinel process and restart a sentinel process .

[email protected]:/usr/local/bin# redis-sentinel myredisconfig/sentinel.conf
423:X 28 Feb 2022 09:22:32.725 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
423:X 28 Feb 2022 09:22:32.725 # Redis version=6.2.6, bits=64, commit=00000000, modified=0, pid=423, just started
423:X 28 Feb 2022 09:22:32.725 # Configuration loaded
423:X 28 Feb 2022 09:22:32.726 * Increased maximum number of open files to 10032 (it was originally set to 1024).
423:X 28 Feb 2022 09:22:32.726 * monotonic clock: POSIX clock_gettime
423:X 28 Feb 2022 09:22:32.726 # Warning: Could not create server TCP listening socket 127.0.0.1:26379: bind: Address already in use
423:X 28 Feb 2022 09:22:32.726 # Failed listening on port 26379 (TCP), aborting.

ps -ef|grep redis // View sentry progress 
kill -9  Sentry progress number 

At this point, the sentinel process can be started normally .

problem 2 Sentinel mode is not responding

This problem is usually caused by the configuration file settings , In the green box below 2 The representative should have 2 Sentinels detected master It's down. , Will really think it's dead , And I have only one sentry here , So this is set as 2, It will always wait for other sentinels to decide , The step of re-election never happened . Change to 1 After the error is solved .

simulation 6379 Downtime

After finishing , The re-election process can be carried out normally

Change hands ,6380 Succeeded in usurping the throne

6379 Resurrection can only be 6380 Of slave

Now let 6379 resurrection , You will find that it is no longer master, It became 6380 Of slave.

problem 3-broken pipe

This problem need not be solved , It takes time to update information ~~~