I was having a issue with twingate getting the same auth.sock error.
So story time.
I was running twingate no issues with 34/35/most of 36 Fedora. So one day just ran into the issue. I could not fix it for anything, uninstall/reinstalled/deleted files etc. So I just create a VM to run twingate in to continue to do my work. Then I decided I would upgrade to Fedora 37. Did that from the cli and it all went well, kinda. I notice upon reboot that it took about 5mins for my network to come up. Now being a upgrade I was thinking maybe a net driver issue. I’m using a HP Pavilion Desktop TP01-2xxx which is kinda of linux supported. Well I was looking around and thought I would run ‘nmcli connection show’ well let me tell you I was surprised. I saw hundreds (yes hundreds) of sdwan0 in connection name. I knew pretty much that was twingate as when I did ‘ip route’ the 100* ip would show up. Well I uninstalled twingate. I then remove hundreds of sdwan0 connections rebooted and bam! My network was fast to connect after I logged in. I then reinstalled twingate. Then I tested it. I did get the “We have detected that another VPN connection is active.” Which is odd as I did ip route and had all 192.168.*.* routes so not sure where it is getting that idea. But I said yes connect anyway. So far at this time the connection is working it is prompting for a long and I have pinged a system for 2 hours with no drops.
So my suggestion is do ‘twingate stop’, uninstall it, do ‘nmcli connection show’ if you see any connections of sdwan0-* remove them (all), then reboot. Reinstall twingate, try to log in.
Good luck,
Jim