00:59:24 voy4g3r2: Advent of code for me was arrogant teens telling me how great they were and how shit I was (unfortunately still a teen at 19 :( ) 00:59:29 this was back when I was 14/15 00:59:45 So I never really understood what AOC is for... other than toxicity 01:02:35 dch: does it current parallel if possible? eg: Say I have 2 codebases... I want them both compiled, two workers can compile one each, but then I want them linked, only one worker can then do this... 01:46:10 AOC is a series of coding exercises, it has nothing to do with alleged community toxicity apart from encouraging light rivalry 01:46:37 sounds like you need to find better friends and take it less seriously 01:47:27 duncan: light rivalry + teenager == bullying :) 01:47:41 and they weren't my friends in the first place... I don't have friends :) 01:47:48 (well not in person that is) 01:47:49 omg, grow up 01:48:26 hey hey that's just my AoC experience... 01:49:18 its why I stopped coding... not AoC specifically but the constant toxicity among developers... sysadmin folks were so much more friendly so I naturally felt more accepted messing with BSD :) 01:50:05 older developers for the most part are helpful and supportive... young ones are not fun... 01:50:42 considering older developers tell jokes about young people (under 25) and flame wars on ML/IRC/whatever... I assume this is a known trend 01:50:59 there was one made at EuroBSDCon 2024 for example. 01:59:25 Maybe AOC would be a good way for me to pick up C again... I abandoned it... wow... its almost been 4 years ago now... feels like just yesterday 02:59:29 the 14.2 betas/rc have been funny about bringing up a fresh/all options listed and unconfigured copy of rc.conf. not sure whats up with that.. I have to delete it all and paste my copy into it in order to have my configurations... ther's no merge /differences being showed.. 05:31:02 It took me one day after 15.0-CURRENT started building packages for me to realize just how long it's been since I've used C/C++. I didn't think I'd be this rusty, buy good grief! With the LLVM 19 changes I am strugglin'! 05:52:29 the cool kids use rust now 05:53:31 aoc is a nice way to pick up a new language. like say, Elixir 05:53:54 i don't need aoc to learn a new language 05:54:53 sure but isn't there that element of shared experience? of everyone working on the same problem and then comparing your work with other people's work? the camaraderie and social feeling? 06:01:11 I've never used AoC before, but the idea doesn't sound bad. Provided the experience polarian had isn't the standard. 06:03:48 bricked my system again updating. maybe I start updating from source for a while. 06:18:39 mrelcee: What are you using to upgrade? 06:19:21 Err... Update. 06:49:04 freebsd-update 06:49:27 shoulda said upgrading 07:00:51 mrelcee: What are you upgrading from and to? (Sorry if you've already mentioned this. My buffer doesn't go that far back at the moment.) 07:01:34 Oh, wait. I see you're going to 14.2. Is that from 14.1 or from 13.x? 07:02:27 14.1-p6 to first 14.2-Beta2/3 and then RC1 tonight. I'll be damned if I try release next. 07:02:49 via freebsd-update anyway 07:03:02 Hrm. I'll run some tests after snapshots (just in case.) 07:03:34 But, I wouldn't be surprised if freebsd-update had a few issues. I *BELIEVE* freebsd-update is being deprecated in favor or pkgBase? 07:03:45 pretty vanilla freebsd zfs on root install 07:04:24 So, maybe there's more focus on that. You could always give that a shot if you wanna really beat yourself up. :) 07:05:32 Where do you think the issues are happening? Just during the config merge or with the base itself? 07:05:41 What are you seeing when you boot? 07:07:25 after it downloads the files, it brings up edits to rc.conf. it presents a vanilla rc.conf file as if I am supposed to verify changes.. i presume the one shipped with 14.2. 07:08:12 I went ahead deleted it all and pasted in a copy of my working copy\ 07:09:55 Seems reasonable as long as the paste goes accordingly. 07:10:28 If you don't paste the /etc/rc.conf, does it work with the default rc.conf? 07:10:43 in my postmortem it is there exacly as is my 14.1-p6 boot env 07:12:40 Where does it seem to be bricking, exactly? Is it panicking during boot or is something just preventing it booting properly/normally due to a setting somewhere? 07:13:53 lots of errors get tossed processing rc.conf. end up booted stuck at single user, press enter for /bin/sh 07:14:49 If you enter single-user mode and edit /etc/rc.conf to the bare minimum, does it boot normally? 07:15:10 Can you paste your /etc/rc.conf and/or maybe a screenshot of the boot errors? 07:15:52 yeah I get root in single user. 07:17:38 Okay. And if you edit /etc/rc.conf in single-user mode to just simple defaults and then exit single-user mode, does it continue to boot? 07:19:19 no 07:20:18 What are the errors when you exit single-user mode after rc.conf is empty(-ish)? 07:25:03 i get a login prompt. it says pam error for any account I try logging into. root or user 07:29:44 Hrm. This kinda sounds like a pwd_mkdb didn't happen or something? 07:30:43 I suspect nothing much after my saving rc.conf happened 07:36:35 Well, the rc.conf merge/save would be during a currently logged in session. The login after a reboot would depend on the pwd database. 07:36:48 this system has been around a long time now. OS has been transplanted between 3 systems going back to 10.x. always updated via freebsd-update. the biggest change to get it working has been the ethernet device name in rc.conf. 07:37:37 If you log into single-user mode, what happens if you run "/usr/sbin/pwd_mkdb -p /etc/master.passwd" and then exit and try to log in again? 07:37:38 very vanilla as far as setup. zfs pool (zroot, storage pool). nfs server, samba server, vm-bhyve. kleep a couple user accounts for running cli tools there and managing the machine. 07:40:31 I cant answer that now. I used beadm to revert iot to the pre-14.2 install state and have it running again 07:41:11 Oh, good! Smart to have a roll-back plan. 07:43:48 If you can log/record your exact steps when you attempt to upgrade again (and it fails,) that would be helpful for filing a PR if this is reproducible. 07:45:04 I'll try to upgrade one of my 14.1-p6 systems tomorrow using freebsd-update when I get the chance. 07:57:12 not sure I am going to have the patience to try again for a whil. I have a year after 14.2 releases.. 07:58:18 and yeah I really dont regret having beadm 08:03:17 might be interesting to try pkgbase if that is working now 12:46:17 One stupid question: can I use the DVD ISO image to do an old legacy MBR installation? 13:45:10 asarch: Do you want to be able to read DVD-IMAGE in a host running old MBR installation - or you want to install freebsd MBR using DVD image? 14:31:28 rtprio: yeah the rust guys are the ones who shat on me and eventually made me not want to touch programming again :P 14:31:40 I am rather partial to C, I just like the language, I think its cool 14:31:47 rust supporters shit on people for using C 14:32:15 I deduced that I couldn't ever be happy because people would never respect my choice of using C, and I don't like writing rust, so I thought fuck programming and went into sysadmin more :P 14:32:52 which is why I attended EuroBSDCon this year, and why I am even in the BSD circles, BSD people were kind and welcoming and I was happier :) 14:39:13 Anyways I guess I will give aoc a go this year... try to forget the painful past :P 14:39:29 If I already have the MBR FreeBSD partition, how could I create the disklabel partitions (eg: /, swap, /usr/home, etc)?: https://imgur.com/a/c65oSvk 14:39:34 ek: don't use me as the standard for everything, I hope for most peoples sake that my experiences are mine alone. 14:55:15 asarch: this might help: http://www.wonkity.com/~wblock/docs/html/disksetup.html#_the_old_standard_mbr 14:56:46 Thank you, thank you very much! :-) 15:10:40 polarian: Rust developers are like the vegans of programming; they shit on everyone who uses anything other than Rust. 15:11:39 remiliascarlet: amen to that! 15:11:46 that happens with webdev, every few years everyone jumps on the latest new thing and shits on people who just code, in native languages. Then a few more years go by and people realize teh bloat ruins everything, and the process repeats endlessly 15:12:18 Seen this happening in real time during the Ruby on Rails craze. 15:12:33 hah that one dissapeared extra fast 15:12:35 And then saw this again during the Node.js craze. 15:12:48 it was all the rage and then you never heard about it again 15:12:59 ugh node.js, aka the worst of both worlds 15:13:13 jquery 15:13:19 To be honest, its not that I hate rust, or people who write rust, but I have bigger issues in life than arguing about what programming language I prefer to write in... and I felt like I was wasting my time programming, hence I dropped it in favour for sysadmin... but obviously my interest in hardware, and low level means I need programming again, and the current debate has reemerged, and its 15:13:21 jquery had more value than most 15:13:21 gotten worse... rust now runs on embedded hardware "decently", rustaceans now say there is no excuse not to use rust for EVERYTHING 15:13:48 but still, it odesn't take much to just learn javascript once you can code in jquery, and just write straight code and skip the 100KB library 15:13:49 web has wasm, backend cli has clap, there is a decent gtk library I heard for rust, and then now it runs on embedded too 15:14:05 The reason why I hate Rust is because of their community's overal attitude. 15:14:08 remiliascarlet: Rust was literally adopted by the web people... 15:14:12 they have the same attitude 15:14:16 "one language to rule them all" 15:14:30 it was javascript, then it was typescript, and now it is rust 15:14:35 And also how they use Cargo for literally everything they want to do. 15:14:44 static linking is a big issue imo 15:14:52 it doesn't fit into the traditional way of distributing software 15:15:06 I dabbled in Linux packaging and the biggest pain in the arse is when compilers static link... 15:15:16 one monolithic binary... 15:15:24 The polar opposite of that is in Go, because in Go you're encouraged to write your own shit using the standard libraries alone. 15:15:47 Rust on the other hand, just load in 3rd party dependencies for literally everything. Same with Node.js. 15:16:13 remiliascarlet: I remember someone saying "rust was not written to replace C, but to replace C++, it was written by mozilla developers who were disillusioned with C++, while go was written by C devs to simply development" 15:16:35 You can write good enough go code which performs well enough for the job faster than you can write decent C 15:17:03 Static linking is actually very useful, since all you need is that 1 binary. Whether libraries break or not, get updated or not, get deleted or not, it doesn't affect the static binary. 15:17:26 C/Go have the same way of thinking though, if you need a feature, you write more code. Rust/C++ has the idea of if you need a feature you implement it into the language 15:17:37 Go was also created because of how long it took to compile C++ code. 15:18:14 a very long time ago, about 21 years ago give or take, I realized that to make the kind of web sites i wanted to make, I was going to have to learn a server side scripting language. So I looked in to my options, and narrowed it down to two: 15:18:19 PHP, and apple WebObjects 15:18:24 thank fucking god I went with PHP 15:19:00 I've seen Go evolve quite a bit. Go 1.22 introduced so many changes to the routing stack, it rendered Gorilla and the other routing dependencies obsolete. 15:19:20 static binaries are a pain in the arse... firstly if a dependency needs updating you need to recompile it, then if a dependency (such as rustls, tls library for rust) is used by many programs it is duplicated over and over and over again... likely different versions of the same library, so say if rustls has a cve, upstream must bump the version lock, and if not some programs might be vulnerable, 15:19:22 some might not... its a fucking headache. Not to mention that static libs dont work in the shared address space afaik, which means that each rustls static link would be loaded into memory separately, instead of shared across processes... more memory use. 15:19:47 GoSox: 21 years ago I thought PHP was the only option for server side scripting, and went with that. 15:19:48 I know loader errors are annoying, and when you have ABI changes, which makes updating dynamic libs difficult 15:20:01 remiliascarlet well there was always perl 15:20:02 lol 15:20:03 but I would still reach for dynamic linking over static linking for ports 15:20:16 You need to recompile a dynamic binary whenever the dynamic library gets updated rather. 15:20:37 Static binaries don't, since the library's code is included into the binary. 15:20:46 remiliascarlet: no, not unless the ABI changes... however I know Linux distros recompile anyways to ensure there is no ABI incompatibility 15:20:48 just to be safe 15:21:01 Only OpenBSD changes ABI's. 15:21:09 but if the ABI remains th same, a library can be dropped in and loaded like normally without recompiling 15:21:14 a linker links symbols, not codeblocks 15:21:44 as long as the linked symbols are the same within two library versions, it should be loaded just fine./ 15:21:57 The reason why binaries break in Arch so often is exactly because libraries get updated, and Arch maintainers forgot to recompile and re-redistribute the binary depending on it. 15:22:09 disclaimer: I am not a low level developer, but I have discussed this with low level developers so I am regurtitating what I took from their knowledge, apologies if there is misconceptions 15:22:25 Unless you installed from AUR of course, because than it's your job to recompile. 15:22:25 remiliascarlet: Arch Linux is a shit show 15:22:49 before I hopped shit to BSD I was part of the Arch Linux community 15:23:00 I am renown for causing trouble :P 15:23:31 I got rather pissed off a lot by packagers disregard for quality, following "KISS" more than standards 15:23:48 the biggest thing which pissed me off most is if something was difficult to package it would be dumped into /opt 15:23:57 with insecure permissions 15:24:10 it worked, why change it? 15:25:09 the [core] repository was always kept to a high standard the [extra] repository mostly to a high standard, with some exceptions of things being left out of date for ages, or things being left broken, but the AUR is a shitshow... 15:25:11 I've been using Unix-like systems for decades, and I still don't get the point of the /opt directory. 15:25:37 Why not just dump it into the paths where all the stuff belongs? 15:25:47 its a landfill directory 15:26:00 its where you dump system-wide shit you couldn't be bothered to install properly 15:26:12 I know that Google really seems to love /opt. 15:26:24 i have used /root/... 15:26:49 They dump the Android Studio, Go, and most other things into /opt by default. 15:26:53 but that's not the point here i guess 15:37:00 remiliascarlet: like google (and Linux) give a fuck about the Unix way of doing things anyways 15:37:22 they piss all over conventions and standards 17:08:44 is there any way to have a poudriere repo "pass through" for certain ports? e.g., if i wanted to setup a local repo following "latest" and config/compile most things, but for big things like firefox have the package pulled from the official latest packages? 17:20:32 polarian: well, /opt is pretty much standard in unix world, solaris documents it as "Root of a subtree for add-on application packages" 17:21:07 I'm struggling to figure out how to use Zfs pull and setting up users with permissions on both servers to do it. 17:22:54 I see a lot of old guides suggesting rsync and theoretical articles discussing ZFS but nothing in terms of this is how you send over snapshots to a different machine via ssh and securely not using ssh keys without passwords on root 17:23:29 very tightly held secret 17:34:58 Is there something I can install from packages to make a FreeBSD box a UPnP client? 17:35:32 When I google on the topic, I keep finding hits mentioning miniupnpd, but that's for a UPnP gateway. 17:36:55 upnp client.. like to open nat ports on your router automatically? 17:42:51 I just installed a fresh FBSD on a machine and kicked of a scrub. It tells me that it repaired >5M – can that happen? I suspect it indicates that the HDDs are dying already 17:43:22 or simply preparing the drive? 17:43:30 how old were the drives? smartctl them to know for sure 17:45:56 Oclair, what do you mean? 17:46:12 rtprio, fresh stuff from 2007 8-) will look into smartmon.. 17:46:32 there should be no inconsistancy requiring scrub on a freshly pressed `zpool create` 17:53:09 silly question...is php 7.4 still available in ports for 14.1 ? 17:56:30 I doubt php74 is still available 17:58:11 ok, my drives have 25000 usage hours already, lol 17:58:19 Maybe ZFS has higher standards and wants to map out questionable blocks than the FreeBSD install formating util 17:58:44 and one is dying. But that's good news – an opportunity to practice pool portation and data recovery in real life 17:58:47 clearly zfs is working as intended with 25000 hours on that medium 18:04:21 I wonder this is still kept in the pkgs then: https://www.freshports.org/www/owncloud/ 18:04:34 it says it requires php80+ but owncloud only works on php 7.4 18:04:41 why not just remove it then ? 18:05:49 Owncloud release 10.14.0 February 26, 2024 18:05:58 kinda risky 18:06:11 Whats wrong with Nextcloud? 18:07:35 so owncloud releases stuff in 2024 for php 7.4 which has been discontinued for 2 years ? 18:07:50 this is next level sillyness 18:08:27 https://central.owncloud.org/t/announcement-owncloud-10-and-php-versions/40251 18:09:08 thats a top pinned topic in their community forums dated 2022 18:09:50 yep, with Docker 18:10:03 that doesn't even matter though..the problem is why is there an owncloud pkg in FreeBSD 18:10:10 when that's completely dysfunctional 18:10:38 request the port be removed due to incompatible dependencies 18:10:55 thats an apolitical way to do it 18:11:07 Nextcloud is very well supported 18:14:01 In fact it was just recently Nextcloud which complained I needed to upgrade php to 8.2 from 8.1 18:15:42 Does anyone have experience using Syncoid on freebsd? https://github.com/jimsalterjrs/sanoid 18:17:38 jemius: next time, apply some ideas from here: https://www.reddit.com/r/DataHoarder/comments/qtbte3/what_programs_do_you_use_for_new_hdd_burnintesting/ 18:26:25 jemius: or, https://www.reddit.com/r/DataHoarder/comments/qtbte3/comment/hkjxx8a/ 18:26:35 that is 18:40:25 Any ideas how to know disk uptime in zfs pool ? Basically I need to know when given disk was inserted and how much io has been performed 18:46:37 ketas, I mean, smarttools usually is the tool of choice, but my results appear surprisingly useless. I read that as "everything alright", but zpool reports repairs https://paste.debian.net/1337746/ 18:49:33 jemius: it reallocates sectors and has them pending, it's fail 18:50:11 checkpoint: i would record when i insert. 18:50:52 checkpoint: i am not so good at io performance for disks though 18:51:07 twss? 18:51:09 :p 18:51:30 zpool iostat -v gives some stats, but it is same for all three disks in my pool (mirror), yet disks were inserted with significant difference in time. Basically I need to figure out which of the disks to replace first. :) 18:51:44 remiliascarlet: home brew folks too 18:51:51 remiliascarlet: they like opt 18:52:00 i recommend keeping disk database 18:52:08 so solve that problem 18:52:20 when was x installed y z 18:52:47 yuripv: can you quote any spec for /opt standard? 18:52:59 ketas, what does "pending sector" mean? 18:53:19 jemius: it can't read or write it 18:53:24 unsure which 18:53:46 that's a big problem already 18:53:56 so does the firmware then typically skip those sectors? 18:54:26 well it retries and put data elsewhere 18:54:35 puts 18:55:15 typically if hdd has issues with sectors, something is failing 18:55:41 hm. But pending sector is listed as category "Old_age", not "Pre-fail". I think I'm too dumb to interpret smarttools correctly 18:55:47 ketas, I was not that smart at the time of pool creation. :) Anyway, disk uptime would solve my problem, is it possible to know when given disk was inserted in the pool ? 18:56:10 history 18:56:20 if it's still in jg 18:56:22 it 18:56:49 zpool historu -il 18:56:51 y 18:57:44 jemius: it already reallocated 26, now 1 is not working apparently 18:57:55 while zfs reports errors 18:58:17 https://i.pinimg.com/originals/01/e7/47/01e7475fbd172a32787fe653b7b7ff65.jpg 18:58:43 there's read and checksum error 18:58:47 bad 18:59:23 disk might not catch error too 19:00:01 ketas, hmm.. yes, I can see 'online' events listed in zfs history. that's at least something I can rely on, thanks. 19:00:11 ah, doesn't matter that much, I mostly set up the machine for the lolz. But I'll search in the basement for another old drive and replace the breaking one :] 19:00:57 basement drives 19:01:00 :p 19:01:25 you could run above tests 19:01:48 then you could replace disks with tested ones 19:01:54 and test removed ones 19:02:13 that'll give you plenty of online disk replacement tasks 19:02:53 record results and now it's all nice 19:02:59 what for? It's broken already. No need to know how broken, except for curiosity 19:03:38 curiosity is good reason too 19:03:49 this is how cat died 19:03:55 not good :p 19:04:56 checkpoint: history is not forever, you could save it if needed 19:05:14 or not if you only needed that date 19:13:18 ketas: how much history is saved ? 19:13:55 ketas: is there any param to set history buffer size ? 19:16:10 i haven't checked 19:16:19 i didn't find iirc 19:18:58 "The log is implemented as a ring buffer. The minimum size is 128 KB. The maximum size is 32 MB.", from here https://docs.oracle.com/cd/E19253-01/819-5461/githh/index.html 19:32:55 how do I replace a broken hard drive? 19:32:56 https://docs.freebsd.org/en/books/handbook/zfs/#zfs-quickstart-recovering-raid-z 19:33:13 is this complete? Can't imagine, I assume one has to format the drive and stuff 19:33:18 *the new drive 19:34:17 hmm 19:34:22 looks half way 19:35:21 I think further down is the relevant section 19:35:22 https://docs.freebsd.org/en/books/handbook/zfs/#zfs-zpool-resilver 19:35:37 basically gpart it 19:35:56 could take ideas from existing disks 19:36:18 and installer scripts are also there 19:36:56 i wrote some helper scripts for that 19:38:12 which do that 19:38:31 of course it assumes you have used it before... 19:39:08 i think some commands are missing in zfs manual 19:39:26 because they are in some other hmm part 19:39:29 adding disks? 19:39:45 partionining and so on 19:40:12 hence it has no exact commands to create table etc 19:40:30 from quick look it isn't there 19:40:40 it focuses mostly on zfs 19:41:10 I know gpart and stuff. Just am not sure about how to feed the new drive into ZFS 19:44:28 zpool replace mypool ada1p3 ada2p3 19:45:34 zpool replace gives 10 results on that page 19:45:42 Seems to require both old and new dev to be online 19:45:57 replacing failed and working drives 19:46:26 well you can replace failed one of course 19:46:50 for boot drives, make sure they actually boot :) 19:47:42 that reminder is there nicely 19:48:22 it won't autoresize by default if you put larger disk in 19:48:47 if you do and resize, pool is permanently large 19:49:20 i think this is still off for anti footshooting 19:50:10 you can use md for playing 19:50:39 small 128m "disks", replace as much as you wish 19:50:56 if you want to test that 19:56:35 Apparently just doing `zpool replace ada0p3 ada1` without partitioning anything works 19:59:16 does 19:59:24 now you can't boot 19:59:28 tho 20:00:11 or you can until remaining disk works and is first 20:00:53 i partition my data disks too 20:01:07 first, size can differ 20:01:14 like in bytes 20:01:29 you also have to remember that this is zfs disk 20:01:43 and booting won't work 20:02:33 if you want to boot off those disks that is 20:02:56 I mean, it's two disks. Would be nice if both are bootable, as before. RAID-2 20:04:23 Oclair, to your question regarding rsnc over ssh, the rsync protocol seems to be the native protocol of the rsync program, not the ssh protocol. Have you tried to do what you want using the rsync protocol? 20:05:57 raid 2? 20:06:59 oh there is raid 2 actually 20:08:35 raid level numbers are weird actually 20:17:49 Oclair: did you get your zfs sync stuff sorted? I would generally either use https://github.com/bellhyve/zelta or syncoid 20:18:01 and use ssh with private/public keys 20:18:17 gaussianblue: rsync has ssh protocol capabilities.. I use it almost every day 20:18:58 gaussianblue: now if its "compiled" in the default binary in the ports/pkg version.. that I cannot say 20:19:16 ( I use i =t primarily between linux boxes atm) 20:19:26 s/i =t/it/ 20:19:37 hmm 20:19:46 thanks gaussianblue and dch, I eventually discovered syncoid https://github.com/jimsalterjrs/sanoid it seems to do what I need from what I can tell 20:20:05 i use rsync instead of zfe send|recv 20:20:43 dch I will check out https://github.com/bellhyve/zelta as well thank you! 20:21:29 but it has drawbacks 20:21:48 It's a little frustrating because this is all included in FreeBSD but I am starting to understand that there needs to be a management tool running the show 20:22:01 apparently i had so much trouble with different datasets that... 20:22:43 now, of course now datasets can't recovered 20:23:08 I have had success with Rsync previously but I am seeking the native zfs functionality because frankly ZFS does things I am in awe of 20:23:12 and one should use snapshots :p 20:24:17 yes I already user beadm and periodic snapshots and those have saved my hide time and time again, but now I want to see if I can archive the snapshots remotely and also replicate the server 20:24:33 *use 20:26:58 apparently recv has issues 20:27:15 i'm unsure how to fix them 20:27:54 Zfs now has that functionality native and the development is getting more and more reliable 20:28:21 gaussianblue: to be pedantic, there's the rsync daemon protocol (rsync://) and the rsync-over-{r,s}sh protocol; not just one 'rsync protocol', though there is some overlap in how the two operate 20:28:35 like what was way to not get tnem mounted on remote machine 20:28:59 the daemon protocol is somewhat horrid and probably shouldn't be used 20:29:23 who usea rsh? 20:29:24 it's convenient if you want to run some ftp dropbox-like thing, though 20:29:28 and rsyncd 20:29:31 in 2024 20:29:44 well, I know of at least one use of rsyncd that's somewhat legit 20:29:52 but it's an internal thing, no expectation of security 20:29:54 public data? 20:30:12 we use it for making /usr/src and /usr/ports and packages available within a cluster 20:30:38 well, I know of at least two uses of rsyncd that's somewhat legit 20:30:51 :-p 20:30:53 can we do 3 20:31:18 maybe dedicated network would justify plaintext 20:31:23 nfs and so on 20:31:28 iscsi 20:32:30 but imagine getting src/ports compromised? 20:32:43 that would be such shitshow 20:34:12 yeah, ideally you have some other mechanism to authenticate the results for a case like that 20:35:13 the first one I'm thinking of is more like a directly-attached device, file transfer of debug stuff 20:35:44 perhaps distfiles would not need it 20:35:56 as ports has sums for it 20:36:12 right, but you still need something to ensure the distinfo weren't tampered with 20:36:33 tho what if ports at fbsd get compromised? 20:36:34 :p 20:38:34 that's not much of a "line" security issue tho 20:38:44 anymore 20:58:19 I just restored a partition table with gpt to my new drive. Worked. But do I need to do something in addition so that the SWAP partition is usable for the system? 21:12:10 jemius: you could/should gpt label it 21:12:12 it has it 21:12:24 but iirc it's not used by label 21:12:43 you could just edit fstab too 21:12:50 where the new disk is 21:33:07 hm, seems `gpart bootcode -b /boot/pmbr -p /boot/gptzfsboot -i 1 ada1` is not sufficient. I get a "ZFS: inconsistent nvlist contents" at boot, followed by "Can't find /boot/zfsloader" and other errors 21:44:26 jemius: swap? that's usually ada*p2 21:44:33 anyone upgraded to 14.2 ? 21:49:38 HER: I haven't done an upgrade... but I am running it though atm.... 21:52:08 Tenkawa: you're upgrading to 14.2 from what release ? 21:52:53 checkpoint: not me.. HER ... I installed it from scratch 21:53:50 Tenkawa: ah, ok.. just want to hear success stories upgrading from 13.x 21:54:06 Yeah I bypassed 13 21:54:18 hey i found a problem, anyone right now if you are running firefox head over to https://github.com ONLY the landing page causes this, your memory will fill and firefox will run out of memoty, and also burn your cpu 21:54:24 checkpoint: from 14.1 to 14.2 21:54:32 want to know if anything will break 21:54:33 its been reproduced on linux by someone else too 21:54:36 I had been doing dev on 15 so I had to reload anyway 21:55:21 polarian: burn the cpu ? brick it ? 21:55:42 HER: exaggeration, it will max one of your cores :P 21:56:01 polarian: real bad spinlock eh? 21:56:11 polarian: nope, seems fine here 21:56:19 HER: platform, browser version please 21:56:33 maybe they fixed it already 21:57:00 freebsd 14.1-RELEASE-p4 firefox 133.0 (amd64) 21:57:45 interesting, 133.0_1,2 for me 21:58:26 polarian: I confirm, landing page of Github consumes 5GB RAM, 13.3-RELEASE-p1, Firefox 128.3.1esr (64-bit) 21:58:27 also HER you are missing security patches, you should run freebsd-update 21:58:37 yikes 21:58:45 -p6 is the latest patchset for 14.1-RELEASE 21:59:21 Microsoft is gonna be sacking people tonight :P 22:00:45 github reports no issues 22:02:09 * checkpoint admits it's the landing page only, directing to homepage (personal account) is ok. Should be a JabaScript issue. 22:03:46 polarian: seems nobody been to their landing page for years :) 22:04:03 checkpoint: your esr version for firefox is quite outdated too, prob should update (latest version is in the port tree) 22:04:45 also yeah I already confirmed its only the landing page, been messing about with it a little bit :P 22:05:53 polarian: I will... soon as the need arises ;) 22:07:16 * checkpoint does "pkg install firefox-esr" anyway to check it with Github issue 22:07:17 you mean, as soon as you are pwned? 22:07:19 xD 22:10:48 hmm... it now consumes 4.5GM only with firefox 128.5.0esr (64-bit), no CPU burnout, that's the progress! 22:11:00 s/GM/GB/ 22:11:56 jemius: observe sample from here: http://ketas.si.pri.ee/misc/zfs-disk-init.sh 22:13:26 ketas, will take a look tomorrow, thx. For today I'll rage-quit I think :) 22:13:49 haahah 22:17:40 the confusion is all HER fault! 22:17:43 nice nick tho 22:21:27 i created that because who needs repeated manual tasks 22:22:04 no efi there tho 22:40:35 angry grunts in basements 22:42:17 what's quite confusing is why the life installer doesn't seem to configure a RAID-2 so that both drives are bootable. That's quite useless, what is someone then to do if the boot drive fails 22:46:45 jemius: are you looking for hardware raid-2? or you meant raid-z? 22:47:17 cybercrypto, no hardware raid. Just a machine with two drives and selecting the ZFS install mode "mirror" 22:47:27 noone does actual raid2 anymore 22:47:45 I am poor. 22:48:04 poor poor boy 22:48:20 but he's a learnie 22:48:29 :) 22:48:52 anyways, even a RAID-10 has to have some solution for the bootloader drive failing 22:49:13 raid 1+0? 22:49:46 wait installer didn't configure both mirror disks as boot? 22:49:59 because last time i tried it did 22:50:14 and i bet it still does 22:50:29 it creates identical setup 22:50:59 note that mirror install also supports different size hdds 22:51:26 just only sized for smallest disk 22:51:47 this machine got 160g+80g at first 22:52:33 apparently i had no other proper machines anymore for tests 22:53:05 jemius: I am asking because you wrote raid-2. 22:53:59 jemius: you want two disks to be redundant or mirrored? I am quite sure you can setup either way. 22:54:28 I can only tell you that I end up with a mirrored ZFS but one of the two doesn't boot 22:56:06 new one? :) 22:56:35 yes. But re-installed 22:56:56 that gpart command is ok btw 22:57:07 but it needs some part table 22:57:55 wait you had it? 22:58:09 unsure what went wrong there then 22:59:15 but yeah, two-disk mirror is somehow called raid1 23:01:48 ketas: Sure thing. I dont care if no one does raid-2 (if you say so). I was replying to a raid-2 question and trying to clarify. Quite simple. 23:04:38 * checkpoint wonders how slow can it be to boot and use FreeBSD from USB flash drive with some basic X11 wm. 23:06:59 i want to make that thing :p 23:07:08 I have to demo yosys synthesis for students at local university. They have only windoze machines at the lab, so I need to boot from somewhere else. 23:07:15 if you take usb3 drive 23:07:21 fast i guess 23:07:34 200m read? 23:07:42 200mb/s 23:09:46 writes can be the problem. anyway, I'm going to try it soon. 23:29:14 checkpoint: I believe if you have at least USB 2.0 to boot from, your system will run reasonably acceptable. I have a usb-drive 2.0 (complete freebsd install) I use eventually on the go (x11/fvwm) 23:31:05 anybody runs wayland with alderlake intel gpu ? 23:31:11 cybercrypto: ok, thanks. 23:33:29 do you need writes tho? 23:33:50 but even they work 23:37:19 ketos: I'll be compiling (synthesizing) some demos, can do it to tmpfs though 23:43:27 sounds fun demo 23:45:05 yeah, I'm popularizing FOSS synthesis tools :) 23:45:41 Hi, how difficult or easy is to take an horpan package and maintain it? is there any onboarding/tutorial/documentation in terms on how to: a) contribute your own package -i.e. you have written a small tool and you want to create a package and propose it as part of the ports- b) take care of an horphan package 23:46:29 s/horpan/orphan/ 23:50:56 what package ?