Mark_LAN

I lied about "Nothing changed". I lost a network component to lightning a few days ago. I am temporarily wedging in some new equipment, which changed my default gateway. Stopped docker and VM's and then updated my unraid gateway address and everything started working. Recommend reviewing your network settings if you see this error!

I am seeing this suddenly, worked yesterday and made no changes. All docker containers are showing "version not available". Any fix? Reboot did not fix it. Plex container is running, but can't see library. Media is there can easily stream from NAS directory. I will continue to poke around....

@Germanus4711 New to me but I only have a light software and random IT background where //is a comment /*so is this*/ and everything else is not. "defaults" being one way and actual comments another is interesting but it does make sense. First time really noticing something done that way, even if it is old school, thanks for the lesson. In the documentation I linked, they stated that performance was the goal. Just not needed for me, a single user environment. 0 for 'worker_processes' disables the multiple executables running the function, but not the function. Docs linked it stays active and is recommend in low ram environments. # for 'min_threads' was defaulting 2 PER WORKER at 800MB of RAM a piece. Supposedly now that I put the min to any number below the max it will actually scale down, for any number of 'worker_processes' BUUUT what matters is I did the recommended changes (removed the '#'s) and re-ran "sudo gitlab-ctl reconfigure" I have reduced about 3GB of ram consumption from over 5GB to just under 2GB for the gitlab-ce docker and maybe 1% or so less of the constant low level CPU chatter gitlab-ce seems to make. That would be a huge difference on a tiny system and was a nice learning experience for me, thanks again for the insight.

Every single value has a '#' in front of it. That's an odd way to me to do a config file. Seems like comments require ## but ill give it a shot. Docs state '0' for "workers" disables the multithreading workers, goes to a single thread for low ram environments, which is more or less what I am chasing. '0' for min also lets the system scale, not sure if that is only in threaded environment or not, either way supposed to scale.

I am trying to learn gitlab in a single user environment (my unraid box in my home LAN). Although I have the resources (for now) I noticed ram usage was high due to 8 puma cluster workers talking up just under 800MB each. Definitely high for my single user needs. I read up on the documentation (here and here) and changed the following in "/etc/gitlab/gitlab.rb" from # puma['worker_processes'] = 2 # puma['min_threads'] = 4 # puma['max_threads'] = 4 to # puma['worker_processes'] = 0 # puma['min_threads'] = 0 # puma['max_threads'] = 4 then ran "gitlab-ctl reconfigure" and also restarted the docker when that failed but I still have 8 workers gobbling up 6 GB of ram. Any tips?