00:16:33 ok...it's back to connecting to the wrong wifi at boot agian 00:39:22 glenny1: I have no idea if this is your problem, but I once had a similar-sounding issue caused by the default wpa_supplicant.conf containing a rule to connect to unsecured networks 00:39:53 wpa_supplicant should be properly configured 00:40:51 I see. Best of luck, then 00:42:38 i don't know what's happening. it's like it's not even reading wpa_supplicant.conf 00:44:16 when you don't know what's happening you're in a bad place, and should turn to logs or tools that could help you understand what's going on under the hood 00:51:24 if i have a bastille jail running a postgresdb, is it "safe" to bastille clone the jail using zfs ? 01:02:44 now dhcp just loops and the system refuses to boot. i'm in single user mode hoping to edit rc.conf but it won't let me 01:04:28 i'm just going to reinstall and try again. i have no idea why things are so broken 01:16:45 reinstall fixed whatever happened 01:45:24 *something* is making the EFI-lua loader on a certain system set "boot_serial" to "YES" (according to the "show" command at the loader prompt when booting) and I can't figure out why. I've combed through the /boot config files and I don't see what could be setting it. 02:06:42 sphex: we derive defaults from the ConIn/ConOut efi vars 02:07:15 `efivar -d 8be4df61-93ca-11d2-aa0d-00e098032b8c-ConIn` may be enlightening 02:16:28 kevans: Ah. Ok. Hmm. I see the loader makes use of EFI variables like this, but not when "boot_serial" gets set somehow... Anyway, is there a way to disable this? 02:17:37 I set "console=efi" (also tried "vidconsole" value) in loader.conf but this doesn't stop "boot_serial" from being set somehow. 02:17:44 sphex: there's a lot magic involved, part of it is in sys/kern for parsing howto flags (RB_SERIAL) and converting them into boot_* vars 02:18:31 But I see "boot_serial" being set directly in the loader (with the "show" command). Hmm something is wrong... 02:18:38 yes, by the magic 02:18:48 what practical problem is this causing? 02:19:59 ohhh this sys/kern file is compiled in the loader too... 02:20:10 kevans: I want it to stop using the serial console. 02:20:25 did you trry boot_serial="NO"? 02:20:39 huh no 02:21:11 hehe. so it's going to be this simple? alright let's try. 02:24:46 my recollection (can't get too distractd from what i'm working on at the second, sorry) is that we rebuild howto flags from boot_* foo right before kernel exec, so it should override the earlier logic and do the right thing 02:28:29 Ok. So yeah it just worked. I just had to set to "NO". Huh I should have tried it before. I let the source code confuse me I think. Alright thanks kevans... 02:29:02 BTW the loader menu is a little bit confused by this. The menu script only checks if boot_serial is set or unset, it doesn't understand "no". 02:48:10 sphex: there are a few other problems with how the menu interacts with console/boot_* vars, haven't had the time to straighten it out 02:48:32 for instance, if you set a console= in loader.conf that doesn't match what we'd infer from ConIn, the menu won't take that into account (iirc) 02:48:38 i began to install a package using pkg but it is pulling from FreeBSD-ports for the package and its dependencies. the FreeBSD documentation says it is best to not use ports and binary packages by pkg on the same system. am i correct here or is the FreeBSD-ports packages different from what the handbook is refering to? 02:55:03 glenny1: are you building packages on your local machine as well.. or just using pkg install for everything on this system? 02:55:15 if the answer is you are NOT building packages, you should be okay. 02:56:09 i don't even know how. i just ran "pkg install librewolf" and it started pulling everything from FreeBSD-ports 02:56:30 you are all good then :) 02:56:39 ok, thanks