Re: tmux, systemd and jack capture

On 2019-04-25 12:39, [email protected] wrote:
> Hey Thomas,
> 
> systemctl --version
> systemd 232
> +PAM +AUDIT +SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP
> +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID +ELFUTILS
> +KMOD +IDN
> 
> From that issue thread I see it's only in 240 - so this is not the
> problem I suppose.
> 
> Cheers
> Jan
> 
> 
> Am 2019-04-25 12:33, schrieb Thomas Brand:
>> On 2019-04-25 11:35, [email protected] wrote:
>>> Hey!
>>> 
>>> I have a hard time setting up a tmux session started with systemd on
>>> raspbian (patchbox OS, realtime kernel) to behaves like as if I
>>> started my script manually.
>> 
>>> Cannot use real-time scheduling (RR/70)(1: Operation not permitted)
>>> JackClient::AcquireSelfRealTime error
>> 
>>> 

It's yet unclear to me if this is related to jack at all.. Can you test 
this without jack? Basically leaving everything out except a script 
started by systemd that tries to lock memory or calls ulimit.

Similar to your tmux approach, with screen, here's a building block:

#!/bin/bash

start_in_detached_screen()
{
         #$1: username
         #$2: screen name
         #$3: command string (without ^M)

         su ${1} -c "screen -d -m -S \"$2\""
         su ${1} -c "screen -S \"$2\" -p 0 -X stuff \"$3\"^M"
}

start_in_detached_screen john_doe "my test" "cd /tmp && ls -ltra"

-> as root: start screen as given user and feed command to screen 
session.
It should work to start jackd as a specific user early in the boot 
process and see what happened by attaching the screen later on.

Greetings
Thomas