[slurm-users] slurmrestd service broken by 22.05.07 update

[Brian Andrus]

I suspect if you delete /var/lib/slurmrestd.socket and then start 
slurmrestd, it will create it as the user you need it to be.

Or just change the owner of it to the slurmrestd owner.

I have been running slurmrestd as a separate user for some time.

Brian Andrus

On 12/28/2022 3:20 PM, Chris Stackpole wrote:
> Greetings,
>
> After updating to 22.05.07 (manually built from source)...
>
> slurmrestd[68695]: fatal: _create_socket: 
> [unix:/var/lib/slurmrestd.socket] Unable to bind UNIX socket: 
> Permission denied
>
> Looking at release notes:
>
> > -- slurmrestd - switch users earlier on startup to avoid sockets being
> > made as root.
>
> OK... So...
>
> $ cat /usr/lib/systemd/system/slurmrestd
> <snip>
> # slurmrestd should not run as root or the slurm user.
> # Please either use the -u and -g options in /etc/sysconfig/slurmrestd 
> <snip>
> # Default to listen on both socket and slurmrestd port
> ExecStart=/usr/sbin/slurmrestd $SLURMRESTD_OPTIONS 
> unix:/var/lib/slurmrestd.socket 0.0.0.0:6820
> <snip>
>
> $ cat /etc/sysconfig/slurmrestd
> SLURMRESTD_OPTIONS="-v -u slurmrestd"
>
> $ ls -ld /var/lib
> drwxr-xr-x. 74 root root 4096 Dec 21 10:55 /var/lib
>
> So then - if I understand correctly, the process needs to run as a 
> non-root user but has to write into a root-only directory to create 
> the unix socket and the only way to adjust this is by editing the 
> service file (which is almost certainly going to be forgotten about by 
> me and replaced by the next version update) and there's no 
> documentation I can find about how we are to configure this 
> chicken-and-egg situation?
>
> Hrm... ¯\_ (ツ)_/¯
>
> I'm not sure if this is a logic bug or if I'm just doing things wrong.
>
> And for note, yes, if I create /var/lib/slurmrestd/, give ownership to 
> slurmrestd user, update the service file, systemctl daemon-reload, and 
> then start the service - it works. But by default, from a fresh build, 
> the process is broken and I haven't found documentation on the 
> recommended method to solve this issue.
>
> Thoughts anyone? Please tell me where I'm going wrong on this.
>
> Thanks!
>