Hello,

did some other test in the meantime:
- on advice of Claes: change RemoveIPC from YES to NO (did also heard some other sad new, not to me to distribute that news)
- changed my nodes to raspi's (cross compiled)

both had no change of proview keeping crashing, same behaivior when auto started

Ron, did you already found some time to try out an headless node?

Ok,
a small update, now I nog exactly when it's happening, but not entirly why.

A small sequence example:
- node reboot
- rt_xtt opened up remotely
- rt_xtt close
--> node crash (or better, proview exits with exexption)

another sequence:
- node reboot
- log in with ssh
- rt_xtt opened up remotly
- rt_xtt close
(no crash at this point)
- exit ssh session
--> node crash

yet another:
- node reboot
- log in through ssh
- exit through ssh
(no crash)
- rt_xtt opened remotely
- rt_xtt closed
--> node crash

There's a link of the last ssh session logging out closing or letting proview failed when in the meantime rt_xtt is opened up. This behaivior is only when proview is auto started and not when starting manually (through script, rt_ini -D or -i or ...). So there must be something happening that links the process when opening up rt_xtt because after clsing the ssh session, everything gets cleaned up and everything exits (with exeption 255).

I assume it has something to do how systemd ties everthing on other system variables and then opening up with the user pwrp gets everything mixed up in the system. Unfortunatly my knowledge doens't reaches far enough to debug this fault but it's a nasty one!

Hello,
small update, tested the behaivior on a "full" installed debian (gnome installed), same result ...

After a lot of reading about the systemd config files, I discovered that there's a seperate config file handling what to do with the processes when logging in. This file is called "logind.conf". All the parameters are dashed out with the standard for RemoveIPC to yes (man logind.conf).

After enabling this parameter and set it to "NO" (also changed the "KillUserProcess" to no) everything seems to work like expected.

Did several reboots, did the same process as before (opening rt_xtt etc.) and everything keeps working: hooray!!
I think I can call this a solution for now

Claes was right again, only the place where to change (probably because the way systemd works changes very much).