00:07:32 I want to enable a swap file after intsalling FreeBSD 14 on AWS. I have dd'ed a 2 GB file from /dev/zero and run `swapon /swap.img`, but it complains that it needs a block device. 00:07:55 I have added swapfile="/swap.img" in /etc/rc.conf 00:11:40 Huh. That swapfile statement in rc.conf seems to have disappeared between ver 13 and 14. 00:13:27 Um. No. After FreeBSD 9. 00:13:37 Okay, I'm a bit behind, it seems. 00:18:04 Okay, a magic md99 line in fstab made it work. 00:18:13 What is that? 00:18:50 Sorry, found the information in the fstab manual. 00:18:51 Thank you guys! 00:26:08 greetings and salutations fine people of Libera Chat \o 01:08:15 dvl, Change is always disruptive. I am rather used to using zgrep because every system requires it. Fortunately I can continue to use zgrep even if not encrypted. 01:10:20 dvl, Why type in bzgrep? That's one extra letter that is not needed. And using it limits the file type needlessly to bz2 files only. zgrep is the same script by the way. 01:13:17 kdehl, I am just going to apply a caution that using a swap file on a zfs file system can deadlock. https://www.schmidp.com/2014/01/05/freebsd-10-does-swap-work-on-a-zvol/ https://github.com/openzfs/zfs/issues/7734 https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1847628 01:13:18 Title: philipp schmid - FreeBSD 10: does SWAP work on a ZVOL? 01:14:23 Remilia, UEFI has been such a road of thorns that this just feels like yet another thorn in our tires. 01:43:58 Will mv command rollback is destination doesnot have enough space? I moved 62 gigs of tar file but running out of space, the file is back to where is was, is it safe to assume the file is still valid? Too big to extract and verify so better to confirm a little bit... 01:44:34 I mean if the destination doesn't have enough space... 01:47:06 that sucks not having the space 01:49:18 yeah... after about 20 minutes it errored with not enough space... forgot to check first, just want to make sure the file is still valid cause it's backup file, kinda crucial.. 01:49:58 for sure 01:52:30 budlight: Thanks for confirming... I'm relived... 01:53:19 budlight is the same person as richard_ 01:54:36 richard_ is on freebsd 01:54:45 in vm 01:56:10 hahah 01:56:29 my main os is debian 01:57:18 reading the bsd handbook 02:05:40 how do i fix sddm and dbus from scrolling down my boot screen i take it out of rc.conf but something keeps putting it back 02:06:31 first time i reboot i forgot to put sysrc infront 02:17:05 cannot help, startx here... 02:18:33 i have that too but i like it to go to the login screen 02:19:18 have you tried others like slim? 02:19:37 no 02:20:11 kde is the only desktop i use 02:21:05 KDE is a Desktop Environment. SDDM, lightdm, slim, xdm are X Display Managers. Meaning that they are separate components from each other. 02:21:33 xdm is rad 02:21:48 xdm is the grandfather of all of the graphical login managers. 02:22:00 * deacon426 smiles 02:22:01 i have try others 02:22:05 I think slim + KDE is also viable... so could give it a try see if it fix the problem 02:22:07 It's pretty plain by today's design standards and so lightdm and slim are the two most popular ones now. 02:23:18 I mma go back and copy that 62 gigs again... 02:23:49 Kinda scared, next time I'll stick with cp instead of mv... 02:24:56 I don't use sddm, KDE's replacement for kdm, but I debbugged through setting the locale with it recently to help someone and discovered that it was not simple. Made me unhappy because it propagates the idea that free software operating system desktops are trouble. 02:26:38 opr, Reading about your mv issue... If the source and destination are on the same dataset/partition/slice then the move will simply move the file and no data copy will happen. 02:27:27 If the move is across datasets/partitions then it will need to copy the file first and then remove the file from the source afterward. If the internal copy reports an error, such as out of space, then it will abort the copy and not do the second part. 02:28:21 But if it is across datasets then I suggest using rsync -a or rsync -t to copy the file first. Especially with 62GB. Then if that works then you can remove the source yourself. Since you know it is a copy then. 02:28:41 what ever got move be mv is lost ant it 02:29:08 by 02:29:32 For that size of a file I would use rsync -aP to include --partial --progress too. So that interruptions will preserve the partial file and be able to resume from where it left off. 02:30:29 Also having --progress there will give some feedback about how long it is going to take. For a lot of files I sometimes like --info=progress2 better. It all depends. 02:30:34 rwp: OH, thanks for the detail, learned a lesson today. It's a system wide backup with rsnapshot, I tarred the whole backup folder than transfer it to my NAS, but selected the wrong disk with not enough space. 02:30:57 I should've automated the whole process with rsync... 02:32:00 Hmm... I probably would have copied the files individually rather than tar'ing them up first. 02:32:41 I was following this blog post: https://subatomicsolutions.org/full-system-backup-and-restore 02:32:42 Title: Full System Backup and Restore 02:32:50 The old school way before rsync was «tar cf - | ssh remote tar xf -» type of thing. But that's not restartable in the case of a network glitch. It would need to do it all again. That's where rsync is better. 02:33:02 VimDiesel: haha You are on it. 02:33:48 rwp: I'll transfer to rsync methodology soon. 02:34:58 I haven't used rsnapshot myself before. There are a lot of backup programs out there. People keep creating new ones! These days I recommend "restic". It's a new kid rising star backup. 02:36:03 opr, Yes the bot VimDiesel is pretty useful to print out titles. But it does get confused if there are multiple URLs on a line and for some combinations of patterns. The failures of the bot are more fun than the successes, honestly. 02:36:45 * rwp is relocating 02:36:56 :D 02:39:26 rwp: a tarpipe can be faster than rsync which is nice for an initial upload. You can resume after network failure with rsync 02:57:11 oh dang, my home is on / 02:57:18 I thought FreeBSD put home in /usr/home? 02:57:24 which O/S am I confused with 02:57:39 * AumShivaya wonders 02:57:49 I guess I haven't used FreeBSD properly since 6 or 7 02:58:25 I *thought* /home was a symlink to /usr/home 02:58:31 so now I'm out of space on home 02:58:36 I guess I could always ..... 02:58:44 Make it so 03:01:39 AumShivaya: changed in 14.0 https://cgit.freebsd.org/src/commit/?id=bbb2d2ce4220 03:01:40 Title: src - FreeBSD source tree 04:02:07 if i make a port can i turn it into a pkg somehow then keep it local and private and pkg install it by passing a path? 04:04:18 polyex: Poudriere docs show how to overlay local packageas. 04:04:21 packages* 04:04:40 polyex: If I am not mistaken, you can setup a pkg repository that is local for you, and as long as you put pkgs in there, you should be able to use pkg to install whatever is in there 04:07:52 anyone make their own tools and maintain a library of private packages? 04:09:20 polyex: https://freebsdfoundation.org/our-work/journal/browser-based-edition/custom-poudriere-packages-in-your-own-repository/ 04:09:22 Title: Custom Poudriere Packages in Your Own Repository – FreeBSD Foundation 04:21:03 zwr, Agreed that opening and closing a godzillian individual files can be a bottle neck for rsync'ing files. Or any kind of file copying. But rsync must stat(2) both files first and that can slow things down. 04:22:31 zwr, With one huge file the one single network stream can be a bottleneck. It can be faster to fracture the file into 100 parts and then transmit each part all in parallel. In a bandwidth sharing situation it will have more shares than a single stream. 04:24:15 People have done all kinds of things over the years to tweak and tune in order to get more speed. 04:39:46 does p;oudriere need root to run? 04:48:37 i can use a bhyve vm for a dedicated poud build machine for my network right? 04:49:49 polyex: sure 04:51:36 polyex: Also, there are not a lot of hits for it, but look at build_as_non_root in the config. 04:51:50 polyex: Also, note that there is #poudriere both on this network and on EFnet. 04:51:54 so poud can run without root? 04:52:34 https://github.com/search?q=repo%3Afreebsd%2Fpoudriere%20build_as_non_root&type=code 04:52:35 Title: Code search results · GitHub 04:52:46 Yeah, it's sort of obscure, but there it is. 04:52:59 really cool ty 04:53:01 Hm, it's the default even. 04:53:14 according to that poudriere.conf.sample 05:55:29 what's cleanvar for? i see it in a rc REQUIRES 05:56:02 i read the script and i see it like cleans var shit but how do i decide if i want my rc to also REQUIRES it? 06:02:03 I have some question, such as: Can libnv be used on other non-BSD systems without Capsicum, and is there a documentation of the pack format of libnv? Also, can BSD be used with non-Unicode text, such as TRON character code? And, can a program work with some GNU functions (such as fopencookie)? 06:03:31 Also, where is the documentation of libuserangel? I could not find it? 06:16:04 the best way to work with non-Unicode text is to convert it to Unicode. And all the BSDs have plenty of GNU functions, yes. Sometimes you can use a library or implement them yourself if they're missing 06:18:12 I disagree. I want to support other encodings without needing to convert them, and supporting the possibility of character sets that may be larger than Unicode or have characters not present in Unicode, etc. I also want to be able to display text without converting it to Unicode (even if it is Unicode text; for security purposes sometimes it is better to display it as ASCII even if it isn't), etc 06:19:38 (Converting to/from Unicode always would be inefficient anyways, as well as inaccurate (although sometimes it is approximately good enough (as a fallback)).) 08:50:52 mason, polyex: no poudriere can't run as non-root itself, as it has to create jails, but it can and does build packages as non-root 09:05:36 Hey, I have problem with ethernet on 14.0-RELEASE, the status is active, I can get an IP with dhclient but still I cannot connect to anything 09:06:45 pinging an IP says that network is down, but em0 is UP and status: active 09:08:07 can you ping your router? 09:11:31 Yeah, I just noticed it is on 192.168.1.1 and not 192.168.0.1, so I was trying wrong IP to access it. However, another device connected through ethernet can access the web no problem 09:12:47 so I have access to the network but no access to the web, weird 09:14:44 bbrtj: netstat -4rn 09:14:52 also check /etc/resolv.conf 09:15:07 verify you have a default gateway 09:15:21 try `traceroute -n 8.8.8.8` or something 09:18:38 Remilia: thank you! I still had 192.168.0.1 in /etc/resolv.conf. Ran `sysrc defaultrouter="192.168.1.1"` and `service routing restart`, and now it works :) 09:21:02 hm, don't you use dhcp? pretty sure you get the default router from dhcp if you don't set defaultrouter in rc.conf 09:30:20 Seems like I'm not, will have to configure it 09:31:04 at least not automatically, as I got IP from dhclient 09:40:51 so you manually used dhclient to get an ip adress? 09:46:40 yeah 09:49:16 not sure if the rc script does additional stuf, but my guess would be that dhclient configures everything correctly if you remove that defaultrouter line and do a service netif restart to just restart everything (no idea what defaultrouter affects), but you can add DHCP into the ifconfig_* line so that the rc script automatically set everything up via dhcp 10:15:23 yes, I will try to set that dhcp in rc.conf later. I will also have to fix network access in jails, as it seems it stopped working. Stuff just broke with my new router 12:49:41 there any way to restrict a user to only running a specific 'service ...' command with sudo? 12:49:48 or is it only per binary? 12:50:56 I have some guy from alibaba trying to connect to my server 1000/sec, I am blocking his IP but pf keeps him in the state table 12:51:00 47.76.35.19:43321 CLOSED:SYN_SENT 12:51:24 I block him with block drop so there's no return 12:53:21 can I even block him from getting registered in the state table ? Eventually he grows to like 50 000 entries 12:54:00 hmm why does something that's drop even build up any state? seems like a drop should be stateless on the server 12:54:35 polyex: you can restict on a matching command string with sudo, which kinda works but is bad design, as you can easily match more than you wanted to allow, so you probably only want to allow exact commands there 12:56:11 hm you mean i can do what i'm asking but it doesn't work great because ppl can like rearrange cli args and get around it? 12:56:57 what would a sudo rule look like to allow not all of service, but only '/usr/sbin/service pf reload'? 12:58:58 looks like doas supports it 12:59:26 permit nopass bobthebuilder as root cmd apt args update 13:02:24 something like hacker10 ALL= (root) /usr/sbin/service pf reload 13:05:27 but that command section also allows wildcards, which for some reason match the command arguments as one string instead of list of strings and * matches space too, so just don't use wildcards in that command part 13:08:41 why ALL=(root) isn't ALL=(ALL) enough? 13:12:51 I just copied some example using the command part and replaced the used command with your command, look the exact syntax up and change it to your needs or change to doas if you want 14:05:51 "I block him with block drop so there's no return" And then he'll return after changing his IP. 14:48:49 nimaje: Thanks for the clarification. Building as non-root is the critical bit. 18:06:48 <_xor> RhodiumToad: I remember you mentioning a while back that SIGABRT is often the result of X reason. Do you remember what X was? (or am I mistaken? There was a coredumped signal that I ran into a while back that I remember you saying is pretty uncommon, and I want to say it was SIGABRT. 18:07:54 assert() ? 18:08:15 <_xor> Oh, that would make sense. 18:08:50 * _xor was just trying to get a terminal image previewer working and it coredumped with SIGABRT 18:09:38 <_xor> Checked the dump with lldb and it looks like it's calling a c++ stdlib method with invalid args. 18:10:28 * _xor needs to rebuild his FreeBSD repo with debug symbols :/ 18:15:02 _xor: uncaught C++ exceptions also caused SIGABRT 19:35:24 most confusing networking problem i've had for a while: pf by default doesn't pass packets with option headers, which includes most IGMP packets, unless you use 'pass ... allow-opts' 20:18:49 unixwitch: iiinteresting… 20:21:51 anyone know if Xorg in FreeBSD 14.0 works with intel iRIS Xe chipsets? When I add: kld_list="i915kms", freebsd locks up solid.. I have to remove that from rc.conf 20:22:34 I have drm-515-kmod-5.15.118_1 which should technically work with that chipset 20:22:37 meena: ask me how many hours i spent debugging why IGMP wasn't working because of this... :-/ 20:23:33 unixwitch: how many hours have you spent debugging why IGMP wasn't working because of this? 20:24:36 that was a rhetorical imperative 20:25:26 I'm not in great form today and struggling more with these kind of things than usual 20:41:24 vgapci0@pci0:0:2:0: class=0x030000 rev=0x0c hdr=0x00 vendor=0x8086 device=0x4628 subvendor=0x1558 subdevice=0x4150 20:42:23 Alder Lake-UP3 GT2 [UHD Graphics] <- any way to make Xorg work with this? kldload i915kms locks up my machine :( 20:47:46 hello. since i updated my bhyve vm to fedora 39, -s 30,xhci,tablet no longer works. the mouse doesn't work at all in the vm anymore 20:49:11 i need to remove -s 30,xhci,tablet from the bhyve command...then mouse works, but without tablet mode, it's not in sync (so it keeps going of screen when i move it around) 20:49:25 Interesting. 20:49:26 has anyone else ran into this and is there any work around? 20:49:45 CorvetteZR1: What host OS version? 20:50:03 seems like someone has/had a similar issue on the forums, but there is no solution posted here: https://forums.freebsd.org/threads/problem-with-mouse-pointer-inside-bhyve-vm.80798/ 20:50:04 Title: bhyve - Problem with mouse pointer inside bhyve vm | The FreeBSD Forums 20:50:13 michaeldexter, freebsd 14-Release 20:50:29 Thank you. 20:51:03 i updated from fedora 36 to 39; i didn't really test it in between. all i know is it was working in 36 but when i started using 39 is when i noticed it 20:51:09 CorvetteZR1: Do try it with several VNC clients. 20:51:17 yes, tigervnc and novnc 20:51:21 Noted. 21:47:05 damn.. https://github.com/freebsd/drm-kmod/issues/252 Oh well.. 21:47:07 Title: Loading firmware on alder lake freezes system · Issue #252 · freebsd/drm-kmod · GitHub 21:47:07 252 – request for a customer id https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=252 21:47:22 Check back on 14.1 or 15.0 :D 21:56:42 as drm-kmod is a port whatever change is needed (likely) don't need to wait for a new freebsd release, but it seems like nobody knows what is needed to be done currently (well, maybe it needs changes to the kernel, but even that most likely works as a -p release and doesn't need to wait for 14.1 or 15.0)