04:01:24 <burie> dch: for the old way, how would one specify ports to be built and put into the image
06:34:03 <dch> burie look in how it’s done in cloudinit modules — https://cgit.freebsd.org/src/tree/release/tools/oci.conf is all that’s needed
07:18:46 <entrop> dch: did you ever get back to looking at https://github.com/berthubert/simplomon
08:34:17 <antranigv> dch educate me. what's that oci.conf?
10:26:24 <uskerine> Hi I am trying to understand a bit better how networking works in FreeBSD. In particular, is it possible to define a virtual interface? Like I would like to have a 10.x.x.x network associated to a virtual interface which is NOT linked/routed to the actual physicial interfaces/networks. I am trying to understand each piece of the configuration/networking
10:26:53 <uskerine> so that interface would be only accessible in the host
10:31:36 <mzar> uskerine: probably you want to create another loopback interface in this case
10:37:57 <uskerine> so I can create a lo1 and assign a network to that
10:38:23 <uskerine> how do you create a new lo interface ?
10:39:00 <uskerine> A loopback interface is a virtual interface that is always up and reachable as long as at least one of the IP interfaces on the switch is operational   <- but it seems linked somehow to the IP interfaces
10:54:59 <ridcully> uskerine: `ifconfig lo create` (or pick some number yourself)
11:03:16 <dch> entrop: no, I found no quick fix for a non-C++ user, so I had to give up
11:03:51 <dch> antranigv: thats the config file for oracle cloud vms. just happens to be one I know well (as the author of it).
11:04:01 <dch> there's an EC2 and other things in there too
11:13:35 <uskerine> is lo always on even if the machine would have no other interfaces?
12:58:50 <burie> dch: How would I specify to use any of the existing or new config in the /src/tree/release/tools directory? It looks like with release.sh I would run release.sh -c /src/tree/release/tools/desired.conf but how would I specify that I want memstick or cdrom target?
13:24:24 <erk> Is there any well supported way of maintaining a port or package outside of the ports tree?
13:25:38 <lw> erk: i just have a local branch of ports with my own ports in, works fine with poudriere
13:26:21 <lw> (i host it in Forgejo, which is in ports)
13:26:47 <lw> the other advantage of that is i can rebuild my packages (e.g. to add a new package) without poudriere re-fetching the ports tree
13:31:25 <dch> uskerine: yes, loopbacks are always there
13:31:43 <dch> erk: yes, just maintain your own ports tree, and rebase occasionally off upstream when you want to.
13:31:55 <dch> you can use overlays (see poudriere docs for this) but IMO its not worth it
13:39:37 <angry_vincent> yes. local branch is way better
13:43:27 <mikewilzn> Currently in the process of migrating my server over to FreeBSD. Very excited! So far all the docs and stuff have been very clear and defaults are sane and stuff
13:43:37 <mikewilzn> Definitely excited to have jails and ZFS
13:46:05 <rtprio> great!
13:49:54 <antranigv> mikewilzn congrats!
13:56:52 <mikewilzn> Thanks! I'm using primarily docker now and definitely ready for a change
14:53:13 <tercaL> mikewilzn: Migrating from which OS?
15:02:46 <mikewilzn> Right now I'm running docker in an Alpine LXC on Proxmox
15:03:00 <mikewilzn> Good timing because I realized I'm still on version 7 and that's going EOL tomorrow
15:17:51 <uskerine> thinking out loud, once you have two separate lo1 interfaces at two hosts, would it be possible to configure a TAN tunnel between both so they are linked?
15:18:24 <uskerine> TAP*
15:26:50 <polarian> I have been looking into wireguard on freebsd and everything uses the userspace wireguard-go, does freebsd not have inkernel support for wireguard?
15:27:09 <mikewilzn> Yo polarian fancy seeing you here
15:27:24 <polarian> mikewilzn: mike from XMPP?
15:27:28 <mikewilzn> Yessir
15:27:37 <polarian> I have been here for ages...
15:27:47 <mikewilzn> I just joined haha
15:27:54 <LapsangS> hi
15:28:08 <polarian> hi?
15:28:19 <LapsangS> yes me too I am a new freebsd user
15:28:48 <LapsangS> or whatever
15:29:19 <polarian> uskerine: LAGG(4) would be better no? I haven't messed with it but it seems like what you are asking for
15:30:05 <polarian> oh wait I completely missed the "two hosts" part, sorry...
15:30:26 <mikewilzn> LapsangS, based
15:30:57 <polarian> loopback should always be local to the device, you wouldn't want to bridge them
15:31:59 <polarian> if you want both hosts to talk to each other, you would want a point to point link, or simply connect to each other over the local network.
15:33:31 <polarian> LapsangS: welcome I guess
15:33:43 <LapsangS> i am just using this system and at slow pace trying to figure out the subtleties of bhyve and zfs
15:34:22 <uskerine> I think lagg servers a different purpose
15:34:28 <uskerine> it is for aggregating links
15:34:34 <LapsangS> for now on, i have the basics working, like X.org and stuff
15:34:36 <uskerine> TAP is a virtual interface to transport L2
15:35:26 <uskerine> I think TUN and TAP are used underneath the (some) VPNs, but not widely used
15:35:41 <uskerine> as standalone tools/mechanisms whatever they are
15:36:37 <polarian> uskerine: I am confused on what you are trying to achieve
15:36:41 <polarian> I thought you wanted link aggregation
15:37:33 <uskerine> I want to understand how TUN/TAP work
15:37:39 <uskerine> and that is not related at all to link aggregation
15:38:09 <uskerine> ultimately I would like to have two subnets, each one at lo1 interfaces, in two different hosts, linked together with a TAP and/or TUN interface
15:38:15 <uskerine> understanding the in and out of that
15:52:53 <scoobybejesus> seems like connecting the hosts requires the nic's to talk, and then you can have routes to the respective loopback devices, NAT'ing out of each of them i guess.  i don't think it makes sense to think of "directly" connecting loopback devices from distinct hosts
15:59:41 <uskerine> I am reading about the TAP interface, still difficult to grasp
16:21:29 <polarian> scoobybejesus: loopback is in its name, its designed for the device to talk to itself, so I am a little confused on the purpose of attempting to join two loopback devices on separate hosts...
16:22:41 <jmnbtslsQE> uskerine: i'm not sure if you will have any luck doing it with tun or tap, i think you need a client program to connect to those interfaces. for your situation, it sounds like you want a gre interface
16:55:11 <uskerine> a client program to connect both interfaces is easier to do, just nc
17:08:52 <jmnbtslsQE> that's not what i mean by connect. i suggest to read more about those interfaces
17:29:17 <polarian> haha there is inkernel support for wireguard, just no documentation other than WG(4)
17:29:18 <rtprio> wait, what are you trying to do ? connect two jails together without the host?
17:30:07 <rtprio> polarian: it does have in kernel support for wireguard
17:30:27 <polarian> I know i found it... :)
17:30:41 <polarian> I asked about an hour ago when I couldn't find anything on it other than people using wireguard-go
17:30:59 <rtprio> the wg-quick package helps, but you probably found that too.
17:31:01 <polarian> and then after reading some history, and also going into the src tree I found the kernel module name and loaded it
17:34:19 <yuripv> i also have wg(8), but may be that's on current only?
17:34:49 <rtprio> no, it's in 14
17:41:45 <kevans> wg(8) is everywhere the kmod is
17:42:16 <kevans> wireguard-tools still incredibly useful for wg-quick and the rc.conf support to make it work somewhat like other platforms
17:43:38 <kevans> there's an rc script floating around in review for base, but I wanted most of it to be split out into a separate script that works kind of like wg-quick and the submitter wasn't really amenable to that / I found them hard to work with
17:47:26 <yuripv> ok, i wasn't following wireguard story at all :)
17:47:58 <polarian> kevans: wg-quick has always been something I considered laziness
17:48:14 <polarian> it doesn't feel like it properly integrates
17:48:25 <polarian> I like how easy wireguard is on openbsd though...
17:49:22 <polarian> although it could be using wg-quick under the hood anyways
17:50:24 <rtprio> it's the same easy everywhere, isn't it
17:50:33 <rtprio> set up the conf file `wg-quick up wg0`
17:51:27 <polarian> it can be done with ifconfig too it appears
17:51:41 <polarian> just its gonna be a long command
17:52:26 <polarian> rtprio: on OpenBSD I did the following: !/usr/local/bin/wg setconf wg0 /etc/wireguard/wg0.conf
17:52:31 <rtprio> not if you want routes. and dns. and the other shit
17:53:07 <polarian> but this works well for OpenBSD as the hostname.if files can run shell commands
17:53:15 <rtprio> that is true
17:53:26 <polarian> I doubt it would work well on FreeBSD
17:53:30 <polarian> rc.conf can be... limiting
17:53:49 <rtprio> yes, i do like hostname.if; i wouldn't mind if freebsd added that
17:53:59 * polarian would highly support freebsd adding it
17:54:22 <polarian> its so much more simple too
17:54:29 <polarian> if you got long configs, you got to cram it all on a single line
18:10:50 <kevans> polarian: I dont have any specific opinions on wg-quick, but I much prefer the model where an rc script is simply a driver for something else that can be used independently to manage
18:11:06 <kevans> i don't really want to review a couple hundred line rc script
18:12:06 <kevans> it's much easier to reason about what the service is providing if the meaty bits aren't conflated right in there
18:12:09 <rtprio> few hundred lines? what sort of config are you running?
18:12:56 <kevans> i don't see how config has any bearing on the rc script's size
18:21:38 <rtprio> but it does
18:23:17 <kevans> ok
19:04:31 <amarks> I've got a Mac Pro from 2007 that has 32-bit EFI with 64-bit amd64 architecture, the EFI loader appears to work, but system hangs trying to load the kernel from 14.1 install img. Any suggestions?
19:24:02 <rtprio> boot with verbose and look for clues?
20:41:38 * CrtxReavr just had a flashback to trying to get FreeBSD installed on a G4 PPC Mac.
20:43:40 <CrtxReavr> Couldn't be done without a special boot floppy from Apple that that could create GTP partitions.
20:44:16 <CrtxReavr> But of course, that floppy image was impossible to find on Apple's website, and none of their support people had any fekking clue what you were talking about.
21:14:13 <amarks> This thing doesn't have a floppy drive. Thanks though.
21:51:28 <CrtxReavr> amarks, 14.x also supports the creation of GPT partiions. . . the FreeBSD versions of the day, did not.
23:29:24 <duncan> wasn't there gparted live CD in 2006? there certainly was in 2008