Just noticed the ever increasing amount of nuci and socat processes since one of the last updates. It just keeps spawning more and more and eats up memory:
root@turris:~# ps |grep nuci|wc -l
181
1 Like
Rebooted less than a day ago and it’s filling up already again. What could this be caused by? I haven’t seen in this in earlier Turris Versions < 3.6.
This is the result of a communication (lost) between Omnia (Turris) and mobile applications Turris. And only when the application in the background checks messages on the router. When you use aplication a normal closing it, ports (nuci, socat) is normally terminated.
Thanks a lot. Not sure if wound consider using socat for this a sane design…
1 Like
As long as this programmers does not fix, I use killall socat if I see that the process is too much (because they occupy memory).
Yeah - https://github.com/CZ-NIC/nuci/blob/c6ee4de9f0c32571b327cd2c76c82110e1599655/src/tls/start is not how a daemon like this should be run. Sorry.
Try again - this is designed for the Turris, with plenty of RAM, right? Why not run something like fastcgi out of Lighttpd?
Experienced today, Omnia gets stuck because updater cannot start (out of memory) and this puts it in non-updatable state (well until restart). Can you guys bump priority on this?
I reported in this thread: Updater failed: Failed to fork command /bin/sh: Out of memory