Huh, that's pretty strange.
When I test it with the standard Debian redis-server package, it
doesn't fail even if I delete the run directory and any pid files...
$ apt-get -y install redis-server;
$ systemctl stop redis-server;
$ rm -rf /var/run/redis*;
$ service start redis-server;
$ redis-cli ping;
PONG
And I can see the pointer to the pid setting in the config... it seems
to be running ok even after it gets deleted?
root@debian-redis-test:~# cat /etc/redis/redis.conf | grep pid
# Note that Redis will write a pid file in /var/run/redis.pid when daemonized.
# When running daemonized, Redis writes a pid file in /var/run/redis.pid by
# default. You can specify a custom pid file location here.
pidfile /var/run/redis/redis-server.pid
root@debian-redis-test:~# ls -l /var/run/redis/redis-server.pid
ls: cannot access /var/run/redis/redis-server.pid: No such file or directory
So missing the config setting should affect it... but it doesnt?
I've got a deploy link to setup a disposable Debian 8 instance that
I'm testing this with if you want to recreate
https://dply.co/b/bGEVS8CF
On 30 November 2016 at 16:11, Chris Lamb wrote:
> Peter Souter wrote:
>
>> This issue doesn't affect the Debian package as it doesn't specify pid
>> location, so it defaults to /var/run/redis.pid, the /var/run dir will
>> always exist.
>
> Are you sure? Under systemd we use:
>
> /var/run/redis/redis-server.pid
>
> (This bug cannot affect the SysVInit script as it calls "mkdir
> /var/run/redis")
>
>
> Regards,
>
> --
> ,''`.
> : :' : Chris Lamb
> `. `'` la...@debian.org / chris-lamb.co.uk
>`-
