00:29:06 seems like jail command should have an option to start jail up even if depend directive isn't satisfied 00:31:27 hm anyone seen RhodiumToad? never here anymore 01:05:55 polyex: I'm not aware of any way to force a jail to start if it has a depend. You can always just temporarily comment out the depend line in jail.conf, though. 01:17:01 doesn't work for my workflow 01:17:05 had to make a dirty workaround 01:18:31 the reason is depend is overloading 2 functions, dep graph, and startup order. would be nice to say something should start before another thing, but not NEED it to start up 01:28:08 polyex: Well, I believe that's the entire point of having a depend. If one won't work without the other, it shouldn't start. Just like any other software dependency. If you'd like it to start even if the depend isn't running, just remove the dependency. 01:28:21 What's the dirty workaround? 01:29:32 again, depend (as in, do not start unless dep is started) is different from startup order sometimes. i want a jail to start after another's started if possible, but still boot if that fails 01:30:01 lemme dig in code 01:44:59 polyex: Ah, I think I see what you're saying. You're using jail_list in rc.conf and depend lines in jail.conf? And the jail_list order isn't always the same? If that's the case, that would be a bug. The man page for rc.conf specifically says jail_list should be the order started and depend lines in jail.conf would be used to override this order. 01:45:13 Maybe open a PR for that. 01:45:25 found a segfault :P 01:45:34 nmap -6 -pn polarian.dev segfaults 01:45:42 it was a typo, should have been nmap -6 -Pn polarian.dev 01:46:13 polarian: Is it just nmap that segfaults? 01:46:26 yes 01:46:49 Well, that's a plus. 01:48:47 ek no sorry i wasn't totally clear. no jail_list because that caused conflicts. just jail enable YES, then depend in jails. so what i want is depend (start jail after and ONLY if other jails are started) but ALSO i want after (order the startup of jail after other jails, but it can start even if other jails fail to start) 01:51:29 oh i use jail parallel start YES too 01:52:44 I've got a weird issue, periodic's daily stuff is staying locked and never completes. This only started a few days ago, I haven't changed anything in the config in months on this machine apart from new IPv6 adressing 01:53:07 ps x | grep perio 01:53:26 returns 10 entries of scripts running and processes marked LOCKED 01:57:59 polyex: Well, I'd guess that's the problem then. If you use jail_list and not jail_parallel_start, I think that'd solve your issue. 01:58:38 can't use jail_list it causes conflicts 01:58:51 "i have 4 jails. A depends on B. B depends on C. and D depends on C. they all start and run fine if i remove the depend directive, but having it in, jail B fails to start because "jail: C already exists". can't depend handle multiple dependants of the same jail?" 02:02:18 polyex: If you have depends for jail C on both B and D jail configs, it fails because C already exists when the second jail tries to start or what? 02:02:58 i guess? but same deps, just removing jail_list, works fine. then add jail parallel start YES, still works fine 02:03:41 vortexx: What does /var/log/{daily,weekly,monthly}.log say? Any errors? 02:04:46 polyex: Thats seems awfully strange (and somewhat backwards from the recommendations via documentation.) I really do think you should file a PR about it. 02:05:16 what doc recommendation are you thinking it contradicts? 02:06:09 polyex: rc.conf (5) and jail.conf (8) 02:06:52 ya but what points 02:09:31 ek: I don't have those logfiles 02:10:04 polyex: rc.conf (5) states that jail_list is the startup order or all jails (unless unlisted) and that "depend" lines in jail.conf would override this order. The jail.conf (8) states that required jails should be started up to the last exec.poststart. 02:10:42 ya so it seems jail_list ISN'T being fully overridden if the depend directives cause problems 02:30:25 Howdy, folks! 02:32:02 polyex: Yeah. That's my thought. Could be a bug. If anything, filing a PR will either bring it to the proper person's attention, or someone will provide you with the correct flow to make it work (hopefully, with explanation so everyone else in the future will also be able to understand.) 02:32:10 Well, for my FreeBSD laptop, i managed to get X.Org set up, as well as getting htop up and going as well. 02:32:33 vortexx: If you kill those periodic processes and manually run a "periodic" command as root, what happens? 02:32:42 Hiya, nicholaus04. 02:32:48 And yes, i even got sudo set up as well. 02:33:55 nicholaus04: Excellent! I hope it's all running smoothly. 02:34:07 Got a total of 198 packages on the system. 02:34:43 ek: Yep, all set up! 02:35:29 Fantastic! 02:35:33 Though, since the laptop uses a Nvidia GPU, i had to get an Nvidia driver from the package repo. 02:36:04 nicholaus04: Was that a problem? 02:36:19 That took some time (Though, not much time.) to get configured. But it's not too bad. 02:36:27 freebsd 02:36:30 Good to hear. 02:36:35 rj1: Ineed. 02:36:39 Indeed, even. 02:36:44 i need freebsd 02:36:52 ineedfbsd 02:36:52 freebsd rules! 02:36:57 Basically, having to silence the ABI warnings which X treated as errors. 02:37:34 And turning off ACPI so the nvidia kernel module would run without it pooping itself. 02:38:33 nicholaus04: I haven't used Nvidia drivers in a while, but I'm glad to hear you got everything sorted out. Did you just follow the documentation that was available? 02:39:09 ek: That, as well as searching for solutions for the rest of the problems. 02:39:36 But other than that, not too bad. 02:39:59 And yes, in terms of X, OpenGL works! 02:40:40 Though, since i tested it with glxgears, i've only proved it works with 1.x API based programs. 02:40:45 nicholaus04: If you ran into issues that weren't documented (and you have additional input you can provide,) you can always submit that info the the documentation site to have it accepted and updated. Might be able to help some others. :) 02:42:31 And yes, i've even enabled (But not tested, nor set up.) linux compatibility. 02:44:41 nicholaus04: Nothing wrong with that. Being as some things are built specifically for Linux and it's libs, sometimes it's required. It's an option for a reason. :) 02:44:42 And yes, like my devuan machine, i'm using OpenBSD's CWM as my X.org WM. 02:45:00 I like CWM. 02:45:11 does freebsd use syncache or syncookies? 02:45:17 i forget the details i used to know 02:46:04 Since i'm literally writing a SDL-ish game framework in C, it would allow me to test for both a native FBSD build, as well as seeing if it works on FBSD's Linux compat. 02:46:22 GoSox: syncookies only (AFAIK) 02:46:48 nicholaus04: Very cool! 02:46:49 All on 100% real hardware! 02:49:21 GoSox: I stand corrected. Looks like sysctl provides options for both syncache and syncookies. 02:49:56 Well, FBSD-14.1-R anyway. Not sure about older versions. 02:50:01 And yes, the laptop is using a i386 based version of FBSD, since that's the sort of CPU it uses. (Not that it's a bad thing in my perspective.) 02:50:49 my server is under essentially constant synflood attack, and my current os has basically zero protection against it. 02:51:04 but between those two options, i should be able to cover my ass on the new machine 02:52:07 GoSox: Bummer! I certainly hope so. You able to utilize something like SSHGuard or Fail2Ban to try and counter it or does that not help due to the exhaustion? 02:52:25 nicholaus04: Nothing wrong with i386. Been a long time since I've used anything 32-bit. 02:53:32 my data center put me behind their anti ddos system 02:53:48 for free, which is nice. but now they want to charge me which is fair but also its super expensive 02:54:03 plus i want my server to be able to take care of its self 02:54:35 pf firewal even has some form of syn flood protection built in, but somehow apple managed to break it in my os version 03:03:47 GoSox: Apple? What do you mean? 03:04:01 its an old mac server 03:04:19 GoSox: Ah. Gotcha'. 05:37:53 Huh. Quick hint: it seems like SBCL conflicts with /etc/periodic/daily/100.clean-disks 05:38:07 Under the mismatched checksums check in the email, sbcl-2.3.9_1,1: missing file /usr/local/lib/sbcl/sbcl.core 05:41:11 Hello, is anyone using Talos Linux over bhyve ? 05:42:14 if we were, we'd be in #linux not #freebsd 05:43:09 do you deploy kubernetes over freebsd ? if so, how ? :-) 05:44:34 i wanna try running kubernetes with freebsd 05:44:38 anyone doing that? 05:44:43 why would you do such a thing? 05:44:52 it sounds.... miserable 05:45:43 isn't it good for controlling deployments? 05:46:54 i want to run freebsd as host, in prod, no matter what. Unless i manage to do deploy k8s cluster with bhyve, i will be forced by superiors to migrate to another host operating system. 05:48:30 L3Fr0g know if k8s can even run on freebsd? 05:49:09 isn't k8s just... .docker containers? 09:02:41 doing an unattended bsdinstall that needs to set up a few jails. they all go fine except 1 of them gets a line in the error log: cannot unmount ... pool or dataset is busy. any clue why? 11:23:30 doing an unattended bsdinstall that needs to set up a few jails. they all go fine except 1 of them gets a line in the error log: cannot unmount ... pool or dataset is busy. any clue why? 12:15:43 Kubernetes is a tooling set to manage (orchestrate) containers (docker) 12:16:22 why can't it work on jails too then 12:18:54 I cannot remember. I've played with it some time ago, but abandoned it. IMHO nobody needs it at home or in small/medium business. 12:19:30 Because whoever concentrates their software development around Docker and/or Kubernetes hates humanity. 12:20:47 what's bad about kubernetes? 12:23:43 when i do a unattended bsdinstall how do i print data into the error log window that shows at the end? 12:28:54 What's bad about installing programs the Unix way? 12:29:33 if you can't just answer without shitting a question back at me i can't take any use from your 'kubernetes obviously bad dur' 12:38:07 Calm down princess... 12:44:02 any idea for: 12:44:02 chouffe% sudo zpool attach zroot mirror-0 da1p3 12:44:02 cannot attach da1p3 to mirror-0: can only attach to mirrors and top-level disks 12:44:03 ? 12:44:50 kubernetes probably has it's use cases, but for deployments with just a few containers it is just over complicated 12:45:12 I just want to add a disk to a mirror zfs vdev 12:45:46 this is what I have https://gist.github.com/silenius/e6e7816369b4581f011512c1dcc30f15 12:45:47 Title: gist:e6e7816369b4581f011512c1dcc30f15 · GitHub 12:46:08 zpool status? 12:46:15 thanks 12:48:13 zpool attach [-f] [-o property=value] 12:48:51 try: zpool attach zroot mirror-0 da0p3 da1p3 12:50:07 ah.. right! 12:50:08 thanks 12:54:36 polyex: I think that Kubernetes can use different "container engines" via a thing called Container Runtime Interface (CRI). It could be something that runs jails or Linux processes under linuxolator, maybe 12:57:35 Just curious, does man (8) pkg-create for anybody else say for the "-a" option: "This option is incompatible with the" and just stops there? 12:58:37 polyex: looks like some is building the tools https://github.com/tnorlin/kubernetes/releases 12:58:38 Title: Releases · tnorlin/kubernetes · GitHub 13:03:42 nice 13:04:03 when i do a unattended bsdinstall how do i print data into the error log window that shows at the end? 13:09:57 polyex: I don't know, something in the $BSDINSTALL_LOG ? 13:11:19 i forgot about that ty! ill check it 13:46:12 hm 'echo $BSDINSTALL_LOG'? 13:46:28 i tried that in my installerconfig script and it just outputted a blank line 13:46:38 but man page says it should have a default value? 13:59:45 polyex: if it is not set, it should have been the file mentioned in the manpage 14:01:25 TommyC: same here 14:05:00 saper you mean if it's empty then it's the default in the man page? 14:06:12 yes, or the default from the actual code running ;) 14:09:04 TommyC: I think this https://github.com/freebsd/pkg/commit/7a5549848b2fcd03a76b7609b707f931b821d790 inserted the text at the wrong place 14:09:05 Title: create: allow to create a package with readable +MANIFEST · freebsd/pkg@7a55498 · GitHub 14:25:55 TommyC: https://github.com/freebsd/pkg/pull/2294 14:25:56 Title: pkg-create(8): fix description of --all by saper · Pull Request #2294 · freebsd/pkg · GitHub 14:26:42 doing an unattended bsdinstall that needs to set up a few jails. in my installerconfig i service jail start jailname, jexec on the jail to do some setup, then service jail stop name to shut it down so i can snapshot the zfs dataset of the jail. works fine but the error log says umount fails because the pool or dataset is busy. so is service jail 14:26:43 stop jailname not cleanly shutting them down? 14:30:40 could happen, any processes left running? 14:30:59 maybe but shouldn't the service jail stop jailname shut them down? 14:31:56 possibly yes, but maybe there is some reason why it didn't 14:35:21 maybe what's happening is: fresh jail is started up in partially configured state, some daemon ran rc services fail to start up and get autorestarted by daemon. then because of that trampoline the jail never finishes shutting down 14:36:00 please check, maybe fstat can help 14:36:53 jail can be as simple as a single process. Also maybe maybe you don't need to start it to configure it. pkg(8) can install packages in jails with the -j option, for example 14:37:14 because i've seen it before that a jail i had that restart bouncing rc daemon issue and when i'd try to shutdown jail host it would get hung waiting for that jail to shutdown 14:37:49 sysrc(8) too 14:38:03 but what about things like adding new user groups, adding users to groups, i gotta start jail for that no? 14:38:22 yeah I've seen things like this too, I'd rather avoid starting jail during the install 14:38:27 no 14:39:01 we can run pw against a jail that's not running? 14:39:27 probably you could use -R option or something like that 14:39:31 saper: cheers 14:43:25 saper so what kinda things would you NEED to start a jail in installerconfig to do? 14:45:09 polyex: well, if you want to initialize a PostgreSQL database cluster inside or things like that 14:45:43 there is also "bsdinstall jail" just learned from https://wiki.freebsd.org/VladimirKrstulja/Guides/Jails 14:45:44 Title: VladimirKrstulja/Guides/Jails - FreeBSD Wiki 14:49:50 I must say bsdinstall looks impressive these days 14:49:55 in unattended bsdinstall installerconfig, i service start a jail, config it with jexec, service stop it, jls confirms it's stopped. but zfs umount ... errors: cannot unmount ... pool or dataset is busy. why pls? 14:50:08 ya unattened install is badass 14:50:14 no more fat fingering an install man 15:14:40 Anyone using devel/bear with 14.1? 15:14:44 Shit, karels@ passed away :( 15:14:45 Does it freak out? 15:14:58 :( 15:19:46 First commit from karels@ I can find was back in August of 1983, and he was last active less than a fourthnight ago. 15:20:09 ffs what a machine 15:20:19 would be cool to see his top 10 commits or smth 15:23:57 He spent a long time being a software engineer involved security at various corporations, which kinda prevented him from working on opensource. 15:23:58 when my unattended bsdinstall is done it shows an error GUI window. is that normal? 15:24:17 wish i coulda learned security from him 16:19:10 in unattended bsdinstall installerconfig, i service start a jail, config it with jexec, service stop it, jls confirms it's stopped. but zfs umount ... errors: cannot unmount ... pool or dataset is busy. why pls? 16:23:16 polyex: try fstat maybe 16:38:16 it can be run on a zfs mountpoint? 16:45:34 Of course you can. 16:47:14 whats the best way to sample all of the more popular GUIs availble for freebsd? does anyone make a 'sampler' with all of them pre-installed? 16:47:17 o/ 16:48:49 I don't think FreeBSD comes with a GUI out of the box, unlike OpenBSD and NetBSD, which come with 3 or 2 window managers. 16:49:24 made a ghostbsd virtual machine and its alright but i suspect theres a better fit 16:49:27 I was under the impression that none of them came with a GUI of any sort, but installing OpenBSD on my laptop shattered that illusion w/r/t that OS 16:50:11 Halian: OpenBSD comes with fvwm (default), cwm, and twm. NetBSD comes with ctwm and twm. 16:51:44 And they both even come with a very minimal graphical login screen too if you enable xenodm on OpenBSD or however it was called again on NetBSD. 17:04:09 Ah 18:48:24 GoSox: pkg install a couple of them and just switch in your .xinitrc 18:48:50 i don't even know whats available 18:49:09 i'm from the land of - not getting to pick the GUI that comes with teh OS :D 18:49:32 * jbo is still not running a DE 18:49:52 once I stopped using DEs my workstation experience became notably less hassle 18:49:53 hassleful? 18:49:56 hasslic? 18:49:59 HASS! 18:51:21 Different strokes for different folks. I'm happy with Xfce. 18:51:26 indeed 18:51:49 exactly on that note: https://forums.freebsd.org/threads/will-freebsd15-have-default-desktop.93784/page-2#post-658712 18:51:50 Title: will freebsd15 have default desktop? | Page 2 | The FreeBSD Forums 18:51:51 it looks like xfce has a dock? is that similar in function to the macos dock? 18:52:12 GoSox: if you have ports, you can expore /usr/ports/x11-wm directory 18:52:27 I am using dwm, but that's not for everyone 18:52:59 I couldn't tell. I've never used a MacOS more recent than 30 years ago. 18:53:02 GoSox, yes, you can have the macos dock experience: https://github.com/helloSystem/hello/blob/master/screenshots/20201229-desktop.png?raw=true 18:53:21 oh i really like that, what is that a picture of exactly? 18:53:41 just a pre-configured FreeBSD: https://hellosystem.github.io/docs/ 18:53:42 Title: hello — helloSystem documentation 18:53:58 jbo i mean what gui is that? and i assume it has a mac theme added on to it? 18:54:14 GoSox, AFAIK it's XFCE with just "look similar enough to MacOS themes" 18:54:42 neat 18:54:45 ill look in to that 18:56:09 vanilla XFCE4 can be customised to have UI similar to MacOS, HelloSystem is maybe derrived from XFCE, but definitelly it's not clean XFCE4 18:56:45 there's xfce4-docklike-plugin in the ports and other goodies 18:56:51 https://hellosystem.github.io/docs/ 18:56:52 Title: hello — helloSystem documentation 18:56:59 i'm not lookign to use a UI thats an exact copy of macos 18:57:05 just one thats more mac like than windows like 18:57:23 If like something original, try cde 19:43:49 Isn't CDE still fully commercial? 19:44:24 It's free software for some years now 19:44:53 I remember when multiple commercial OSes shipped it and thinking it looked so cool. 19:45:10 I remember it was quite heavy back in the days and today it feels very lightweight 19:45:27 But then when you attempt to daily-drive it, you realize it's a horribly buggy POS, and can't imagine that anyone would willingly pay for it. 19:45:27 i had CDE for Linux on a laptop many years ago (not the Caldera one, a different port) 19:46:37 I used it on AIX and it was very stable for me. 19:46:57 Solaris, AIX, SCO Unixware, HP/UX. . . 19:47:07 Trying to think what other OSes I used it on. 19:47:38 It wasnt' so much instable. . . it just didn't work in so many ways. 19:47:52 Cy Schubert maintains a very polished CDE port for FreeBSD 19:48:04 Like if you accidentaly moved that button bar off center, there was to way to get it centered again. 19:48:16 What's it called? 19:48:45 The FVWM2 that used to ship with BSDi looked really nice. 19:48:55 x11/cde x11/cde-devel 19:49:24 unicode support got better 20:33:03 Ah, CDE ... brings back memories... I think the last time I used it was back in uni ... so ~2006 at latest, they still had some HP-UX workstations labs 20:46:23 I think they called it VUE or something 20:53:37 VUE is something else, that's the one that came before CDE 20:55:43 vue,openwindows/libxm xview,sun news, sun did couple, 20:56:05 although interestingly wikipedia claims CDE was largely based on VUE which is interesting 20:56:08 motif was $ 20:56:37 CDE meh...hp,sun,ibm settled on 1 thing, windownt ate their lunch 20:58:53 comp.windows.open-look and alt.toolkits.xview 20:58:59 usenet 20:59:05 I loved VUE 20:59:17 And yes CDE is very much based on VUE in the visual sense 20:59:42 IBM brought in the "boring business" look 20:59:47 And Sun I don't know what they brought 21:00:03 VUE was much fruitier 21:00:36 And yes it was all based on motif like pretty much everything those days 21:00:56 no mwm was not same 21:01:22 libmotif/mwm was pay to play. 21:01:55 till lesstiff anyway 21:01:58 oh I don't remember mwm 21:02:01 But fvwm 21:02:11 https://en.wikipedia.org/wiki/LessTif 21:02:12 Title: LessTif - Wikipedia 21:02:24 * wildeboskat still has a HP-UX 10.20 box with VUE <3 21:02:46 yeah but do you have bootable install tapes? 21:02:50 ignite ux 21:03:14 hp also charged for raid1, mirror ux 21:03:26 you want tmpfs..yeah pay for that also 21:03:45 hp-ux was crazy 21:04:34 I don't have tapes no 21:04:36 posix sh for shell, compiling kernels for ddt tape drives..one first things i did 21:04:44 I had a J210XC that did have tapes 21:04:53 But I sodl it because it was a huge tank of a machine 21:04:58 i didnt have to compile kernel on solaris, but hp-ux compile 21:05:04 I do have CDs and an old SCSI CD Drive 21:05:21 Yeah and HP-UX also came with an amazing management tool, SAM 21:05:22 hp c8000 on ebay NOS pa-risc 21:05:32 new old stock! 21:05:35 I think I have a 700/80 or something 21:05:39 Also PA-RISC 21:05:49 Itanium only runs HP-UX 11i 21:05:50 yeah.. 21:06:09 They were throwing those itanium boxes out at work and it was sad to let them go 21:06:14 But they are so huge 21:06:30 And they run nothing anymore, not even linux 21:06:39 i had rp8400/superdomes, l,d,k,v2600 was top500 supercomputer 21:06:52 Yes we had them too at work! 21:06:55 v2600 128cpu/128GB = 4 cabinets 21:06:55 The superdomes 21:07:02 in the 90's 21:07:05 We had an old beige PA-RISC one and a black itanium one 21:07:13 Both just 1 cabinet 21:07:29 And all talking to the SAN 21:07:29 ssp was b180l for v2600 21:07:38 b180l nice box 21:07:39 We had them until way into the 2010s by the way 21:07:54 yeah i walked away in 2006 21:08:00 In fact they used to come and ask me questions because there were very few people that knew HP-UX 21:08:19 I wasn't technically part of IT but I was in the friend zone lol 21:08:35 That was handy when things were being thrown out <3 21:08:45 my boxes interfaced with ibm system Z and sun e25k real nice boxes 21:08:55 Ohh I never worked with system Z 21:08:58 i wish i could have taken photo 21:09:06 We didn't have any of that 21:09:11 system Z and sun e25k sitting next to each other 21:09:11 We had one AS400 at one point 21:09:18 But that was tiny stuff 21:10:02 yeah system36/39 before as400, as400 was virtualizing 21:10:23 ibm was ahead of everybody on virtual machine foo back then 21:10:31 yes 21:10:38 Although they did most of it in hardware IIRC 21:10:41 but assholes 21:10:43 yes 21:10:45 jcl/rpg2 21:10:48 The high priests of IBM 21:11:06 i was pissed at ibm before the as400 came out 21:11:22 I feel like somehow those days were more adventurous 21:11:27 that how they did things high priests bullshit 21:11:31 Different architectures, OSes, always somethign to explore 21:11:42 Yes, if you were not blessed you would not get to touch anything lol 21:13:47 ibm holding the crown jewels close, $$$ no opensauce/freesoftware with them 21:14:02 least they dumped 1billion into linux to fight off m$ 21:15:23 https://www.businessinsider.in/IBM-Vows-To-Spend-1-Billion-To-Promote-An-Operating-System-That-Competes-With-Windows/articleshow/25697930.cms 21:16:01 This is the second time that IBM coughed up $1 billion to promote Linux. The first was way back in 2000, when Linux was a fledgling operating system just finding its way into enterprise data centers and beginning to threaten Microsoft. 21:16:47 dell emc redhat box crapping the bed in 2004 classic...compared to my sun hp boxes 21:21:04 Netflix freebsd cdn boxen making #freebsd better! 21:43:58 Howdy, folks! 21:46:14 Well, i managed to goof with producing sound via /dev/dsp via the shell. 21:47:41 Basically, i did this: yes (WHATEVER)>/dev/dsp 21:47:57 hey folks 21:47:58 I'm having issues with AMD GPU. it's a MacBook Pro 2015 21:47:58 the Intel GPU one is working fine 21:48:02 this one, not so much 21:48:06 who can help me debug? 21:48:11 I instaled drm-kmod 21:48:16 and drm-515-kmod 21:48:24 when I do startx, the screen freezes 21:48:26 but I can still SSH in 21:48:32 And for the safety of your ears, don't do: cat /dev/random>/dev/dsp. 21:51:31 here's what I get 21:51:32 RADEON(0): [drm] Failed to open DRM device for pci:0000:01:00.0: No such file or directory 21:51:37 RADEON(0): Kernel modesetting setup failed 21:51:49 and it ends with: Screen(s) found, but none have a usable configuration 21:56:29 And yes, i know FBSD uses OSS as it's default sound output. 22:26:59 fstat can be run on a zfs mountpoint? 22:46:24 in unattended bsdinstall installerconfig, i service start a jail, config it with jexec, service stop it, jls confirms it's stopped. but zfs umount ... errors: cannot unmount ... pool or dataset is busy. why pls? 22:52:38 question: why do you stop the jail during the install? 22:53:08 so i can snapshot it 22:53:21 its zfs fs 22:57:36 polyex by any chance devfs is still running? 22:58:01 in the jail? 22:58:47 still mounted, rather than still running 22:58:50 but good idea 22:58:55 polyex yes. 22:59:18 polyex do something like 'mount | grep path/to/the/jail' to see what else is mounted 22:59:21 how do i check? 22:59:25 ok 22:59:30 polyex lemme know 22:59:37 checking now brb 22:59:52 btw fstat /path/to/jail sohowed nothing 22:59:53 k brb 23:01:14 polyex fstat shows files opened by processes, not mountpoints of the kernel. 23:01:25 polyex how did you create the jail btw? 23:02:43 polyex if devfs is not running, try "zfs unmount -f yourpool/datasetzfs". No need to bsdinstall. Just fetch the base, extract and make a snapshot. 23:02:54 install running, will have info soon. i created the jail by zfs clone template into its own mountpoint. then i cp my jail's config files into its dir then service jail start myjail 23:03:14 polyex if that's all you want to do, please have a look at https://jailer.dev/ 23:03:15 Title: jailer 23:04:19 i'll check it out after i fix this thing 23:05:21 devnull well i tried zfs umount /zroot/path/to/jail but it said pool or dataset is busy. so you're thinking try -f to force it? 23:07:52 ok it says devfs in addition to jail's mountpoint 23:07:59 zfs, local, noatime, nfsv4ac 23:08:30 antranigv looks like devfs is mounted for jali ya 23:08:39 polyex unmount that first 23:08:57 zfs umount /zroot/path/to/jail/dev ? 23:09:15 polyex it's not ZFS. just `umount /path/to/jail/dev` 23:09:49 and path is FS path not zfs mountpoint path? 23:10:13 polyex indeed. your Jail path is a ZFS path, but the /dev inside of it is just a devfs mount. 23:10:36 do umount then do mount grep line again to see if it's gone? 23:10:50 polyex indeed 23:11:39 install running 23:26:13 OK 23:27:04 umount line did the trick! devfs mount is gone. then error log no longer shows when bsdinstall is done!!!! 23:27:14 tyvm antranigv! 23:27:30 polyex congrats 23:27:33 so what i learned is that if my jail uses devfs, i need to unmount that after shutting jail down 23:27:43 polyex now use jailer :) it has all of these issues fixed 23:27:57 does jailer also upstream the fixes into base? 23:28:09 in jailer it's as simple as `jailer image fetch && jailer create jailname` 23:28:32 polyex thanks to Jailer (and other Jail management tools) there has been two fixed and a feature in upstream! 23:29:00 that's so great. i'll look at jailer for sure. know what it's coded in? 23:29:26 polyex POSIX Shell with FreeBSD extensions. 23:30:50 nice 23:35:23 did you make it? 23:36:28 polyex yes 23:36:41 polyex the history is at the bottom of the page 23:44:52 you ever run into prob that vnet jails can't do mac_portacl? 23:44:59 would be great if that was added 23:45:16 so we could allow a regular jail user to bind to a specific low port 23:48:49 polyex jails can do that, AFAIK 23:48:55 nope 23:48:59 testing now 23:49:48 last tested on 13.3 fwiw 23:49:57 hopefully you're testing 14.1 in case it works now 23:52:00 polyex tested on 13.2 and 14.0 23:52:02 root@e0fc9570:~ # sysctl security.jail.jailed ; nc -l 80 23:52:02 security.jail.jailed: 1 23:52:02 hello from the other side 23:52:09 polyex works fine 23:52:42 oh wait 23:52:45 this not VNET 23:52:45 sorry 23:52:47 1 sec