Why does the redis socket drop after a server reboot/upgrade?

Question

linux

Hello. Faced such a problem. Very often you have to expand and upgrade servers, and it turns out that after a reboot, the radish socket constantly falls. It hangs in tmp (the path to the socket is specified in the radish config), a non-private tmp line has been added to the radish service, but when you start the Forever (or another utility that we use radish) - the radish is dead, that is, the socket is, as it were, but it's like he doesn't exist.
Nobody faced such a problem?
I just can’t understand, the problem is in the radish itself or in the code of the developers, that each time you have to rebuild the radish and its configs.
P.s the memcache socket, etc. continue to work normally

Answer 1

Understood. After a reboot, some kind of trait does not clear the socket itself from the directory specified in the unixsocket. That's why he can't connect.
The problem is not in the radish, but in the programs that constantly try to create a socket in the tmp with a script