00:02:25 rwp, are you starting all 4 cloudinit rc.d scripts? or just `cloudinit_enable=YES` ? 00:02:48 it seems here this only works if cloudinitlocal, cloudinit, cloudconfig, cloudfinal are all started 00:03:04 which IIRC wasn't the case ~ 1 year ago when I did the OCI cloud setup 00:42:01 dch, Sorry but so far I only have Debian systems at Amazon. 00:42:06 So I feel I could talk about the Amazon side of the setup since that is the same. But I don't know about the FreeBSD side of the setup. 00:42:36 However I am in the process of migrating things with the goal of changing the systems I admin there over to FreeBSD. 00:42:42 ok, this is very freebsd specific. the port docs just use cloudinit_enable=YES but this doesn't seem to run the appropriate stages, although it may have done this in the past 00:42:49 its 2am here so I'll check it .. later 00:42:53 So at some point here maybe not too far in the future I will be working through FreeBSD systems there. 00:44:35 Sorry but I don't have first hand experience with the FreeBSD rc script startup for cloudinit. 00:44:57 np, I will acquire that and you can ask me next time ;) 00:45:13 For me the way I do things I just need it to fetch a single script of mine from another server of mine sideways across the network and then run it. 00:45:29 And after that starting point I do all of the rest myself in the script. 00:45:52 I really only need cloudinit for the very first provisioning and then never again. 00:46:23 yeah, this is the bit thats broked. TBH ec2-scripts are far simpler and less WTF-prone. 00:46:29 http://www.daemonology.net/blog/2013-12-09-FreeBSD-EC2-configinit.html & www.daemonology.net/blog/2018-12-26-the-many-ways-to-launch-FreeBSD-in-EC2.html 00:46:30 Title: Introducing configinit 00:49:09 configinit looks to be similar to but not identical to the cloudinit I am familiar with. But being designed for Amazon use it seems like it should be functional enough. 00:50:08 configinit looks to be a domain specific language reading a data file to edit FreeBSD config files specifically. 00:50:46 Whereas I prefer a script where I can call *my own domain specific* setup process. But it is my process instead of another's process. :-) 00:51:03 Anyway... Get a good night of sleep and I am sure things will work better in the morning! Good night! 00:57:22 Hi, how can i block a package from being installed at all? I tried to lock the package but when i install something that has it as a dependency i have to install it and force delete it. 01:06:09 rubynt, I don't know but... I think you might have to create an equivalent package that's empty and install it to satisfy the dependency. 01:06:24 Because I don't think you can lock a package that is not installed to keep it from being installed. 01:30:13 rwp, thanks 01:34:04 Hopefully someone else who is more knowledgeable will know and jump in. 01:36:35 poudriere and custom pkg repos 01:36:48 or live with it :) 01:37:12 mmm, poudriere 01:39:12 cant really tell a pkg manager to not install a pkg when it requires it fomr where 01:39:21 pkg is awesome but still a pkg manager 01:39:56 another option would be to create flavors for said port 01:40:09 subpackages! 01:40:10 or stop caring as HD are large enough that 15 meg pkg shouldnt hurt 01:40:12 aaaaany day now. 01:40:34 I'm being a butt, but subpackages _are_ pretty much the solution. 01:40:44 debdrup, welcome to my world 01:40:52 What, being a butt? :P 01:41:01 amongst other things 01:41:16 if you want custom pkg poudriere 01:41:36 kinda overkill for just one pkg or set of packages but 01:41:38 it works 01:47:42 rwp, what port is it 01:47:53 pulseaudio 01:55:40 cpet, Not me. Ask rubynt! :-) 01:56:19 hrm 01:57:46 I don't know what package rubynt was trying to block. Names were not mentioned. 01:59:12 i get upset when I install git 01:59:20 then I look at my 1TB SSD and go bleh 02:45:46 rubynt: how would you control all behavior of arbitrary packages in the future? what if you just download some complicated package and it internally wants to have its own package X you thought you prevented after you get your hand dirty? 04:01:09 mictty, either all pkg or all custom not both 04:01:35 the only exception is compilg a port without and option changes 04:10:06 portmaster was resurected ? 04:37:46 cpet: I don't know how people manage packages but I see sometimes a large package has its own package such as python internally for compactibility. 04:38:39 always a way to use the system package 04:38:47 cpet: or I didn't understand what rubynt talked about...? 04:39:03 i dont know 04:39:19 right now only thing keeping me going is the beer 04:39:21 ;/ 04:42:02 cpet: beer-driven development 04:42:43 think the only thing ive done is update jailkit 04:44:26 cpet: updating? good enough. take care. 04:44:50 doesnt have a maintainer 04:44:58 and I was going to use it then no 09:24:36 dch: seems that incompatibility errors are gone, if somehow we see something else I'll update 09:24:51 thanks 11:22:54 I'd like to hear how people, who want non-installed packages blocked, want dependency chains solved. 11:24:40 I'd be surprised if that isn't a non-deterministic polonomial. 11:48:59 debdrup, not sure why it would be NP unless solving the dependency chain and load order already is. Just check each dependency against the "never install" list, and if it's there, fail the install, upgrade, or whatever it is the user requested. 11:50:27 (I don't see a way to handle it other than failing the operation.) 11:51:44 According to Titus Winters (a google dev bigwig) dependency management is the hardest problem in computer science 11:54:58 I'm pretty sure that's been known since the 60s.. 11:56:15 * shiroyasha thinks dependency management belongs to the version control layer, and not the language tooling layer. 11:57:13 (Tangential comment.) 11:57:41 Maybe? I suspect it depends critically on how the utility is supposed to handle such constraints, eg fail vs. try to find ways around them. 11:58:29 (I wasn't discussing the larger problem, which I know nothing about.) 12:08:51 I'm not sure you can just throw version control at every system and expect it to solve dependency chains for all use-cases. 12:10:04 Aside from the practical considerations of how much code you'd end up adding to basically every execution which would inevitably slow down things, even just a repo like that of FreeBSD is big enough that git isn't exactly what one might consider fast on a system equipped with an NVMe SSD and 16GB of memory. 12:12:53 * V_PauAmma_V puts a pancake on his head and stares blankly. :-) 12:13:19 I thought you were a crab, not a bunny? 12:13:39 If Snape can do it, so can I. :-) 12:13:58 True, no need to let bunnies keep you down. 12:14:46 Anyway, groceries run. 12:16:44 I'm thinking of moving from Ubtuntu/Debian to FreeBSD. What is the upgrade process like? Do you need to wipe clean and recreate the OS and then install packages? Or is there some upgrade functionality built into the OS? 12:17:21 To be clear I'm asking about after I move from linux to freebsd, what will the upgrade process look like from version to version 12:23:47 FNAShinobi: there's freebsd-update utility to help you with that 12:24:48 More about the topic is here: https://docs.freebsd.org/en/books/handbook/cutting-edge/#updating-upgrading-freebsdupdate 12:24:49 Title: Chapter 25. Updating and Upgrading FreeBSD | FreeBSD Documentation Portal 12:35:47 Is there any potential gotchas that I may have to worry about? E.g. migrating my linux zfs pools to bsd 13:07:56 I never used OpenZFS on Linux, but as long as you're using the same OpenZFS on Linux and FreeBSD (which means FreeBSD 13.x as 12.x uses a different version), the only likely problems would be Linux-specific features (if any - again, I never tried it on Linux). If the version in base FreeBSD doesn't work for you, you can try the port or package, which was updated less than 3 months ago. 13:12:31 Thanks 14:15:05 hi 15:09:21 debdrup: its almost always pulse audio 15:43:25 pings meena 16:54:33 pulseaudio is garbage. pipewire is slightly cleaner garbage. 16:54:47 the former seems to work better. 17:18:57 i agree. although i use oss 17:23:14 Hi, I am having problems using nvidia driver in order to start Xorg. I installed nvidia-driver-340, my graphic card is a Nvidia 320M and the laptop which I am using is a MacbookPro 7,1 Mid-2010 17:23:18 https://bsd.to/wEsx 17:23:19 Title: dpaste/wEsx (Plain Text) 17:24:25 sedzcat: did you try the -ignoreABI option? 17:24:31 Yesterday I managed to copy the system to this macbook pro using prime-list via pkg 17:24:47 Where do I put -ignoreABI? 17:26:44 xorg.conf 17:27:01 Ok, I'll try 17:27:48 Section "ServerFlags" 17:27:48 Option "IgnoreABI" "true" 17:27:49 EndSection 17:28:56 (I have a 15" and 17" mid-2010 mbp) 17:30:56 Wow new thing here, it reboots the system when I startx 17:33:58 Endless reboots since I have slim login manager enabled 17:35:06 ouch 17:35:19 (I dont use mine with freebsd fwiw) 17:35:31 theres some chatter on posts about using an old nv driver 17:35:36 good luck 17:45:04 dch: pong 17:45:20 dch: I need to establish a routine… 19:11:18 excellent the more people leave freebsd, the more there is left for me!!!! 19:15:33 not true 19:15:36 it's not a pie 19:28:29 is it a cake???!?!? 19:29:41 BSD == big, spongy, decorated (cakes!) 19:56:14 The cake is a lie. 20:15:52 the marginal cost of copying is low 20:19:40 life without cake is low 20:27:34 meena: gnop 20:27:49 Big Spongy Dessert 20:29:48 sedzcat: you can try the older driver-304 but your best bet is to look up on the NVIDIA site to see which driver version matches, and try that 20:31:49 FNAShinobi: the major thing to watchout for is that xattr is a common default on ZoL, but is IIRC only supported in newer FreeBSD ZFS versions 21:13:09 dch it is really driver-340, that's what said the nvidia website