Redis fault handling process is already running or crashed

Today online a test library redis, add auth password verification, I want to restart redis, do not restart does not take effect. No matter what happens, NOAUTH Authentication is required. redis-cli -a 123456 shutdown unexpectedly did not use, in anger, forced shutdown, kill -9 redis pid number.

And then it starts. The requested URL/var/run/redis_6379.pid exists, process is already running or crashed.

Since it was a test library, the data in it was not too important, and the solution was rather crude.

I'll just back up the dump.rdb file to another directory,

Then delete the pid number of (rm -rf /var/run/redis_6379.pid) redis

Start directly [root@localhost6379]# /etc/init.d/redis_6379 start

It's working.

So the question is, what if the data is more important? Just copy the dump.rdb file back?

no kill no die