00:00:35 city in Canada, guessing on _ca? :) 00:01:44 In Ottawa (was getting dis- & reconnecting too much also today, so gave up) 01:02:45 Possibly off topic, in case any one uses|used "No-IP" & thus TrustCor : https://www.washingtonpost.com/technology/2022/11/08/trustcor-internet-addresses-government-connections/ 01:50:01 https://freebsdfoundation.org/our-work/journal/browser-based-edition/ 01:50:05 Title: Browser-Based Edition | FreeBSD Foundation 01:53:19 <_xor> If I want to boot a bhyve Windows guest using MBR and not UEFI, what's the right bootrom to use? 01:58:35 trench, Thanks. 01:59:16 meena, I am ready to sign your petition of ssh being enabled in a VM 02:05:13 _xor, the default bhyve one 02:08:15 Cute, FreeBSD Journal, for I was not expecting an article to be presented as a PDF file 02:08:28 ... in "browser-based-edition" 02:29:50 Previously when I ran `pkg upgrade [package name]` and the package was not installed, it would give me a message that said the package was not installed, nothing to do. Now pkg prompts me to install the package just like I invoked the command with `pkg install ___`. Is this the intended behaviour? 02:31:56 If I'm reading the manpage correctly, it says it should skip packages if not installed, right? 02:33:22 I can confirm 02:34:08 (... the offer to install a package when trying to upgrade one that has not been installed) 02:36:22 parv, thanks for the confirmation. 02:36:28 Manual page of "pkg-upgrade(8)" -- Moreover, pkg upgrade will not install new packages, except as required to fulfil dependencies of the packages listed on the command line -- is severely out of date from the behaviour (on stable/13) 02:37:47 Right, that's what I read too. This new behaviour is messing with my invocation of `pkg upgrade` from Ansible. I only want it to target systems that have the packages installed. 02:38:25 hackershack, Thanks for bringing that to attention. Else I likely would|might have pressed "Enter" without reading the message closely 02:38:25 I'll open a bug report. Thanks for the second set of eyes. 07:35:03 I am running 12.3 on powerpc and having trouble doing a buildworld of 13.1, I suspect because the compiler is still GCC but Clang is expected, does anyone know how to deal with that? 08:53:54 but why listing packages there instead of using just pkg upgrade without listing packages? that should upgrade all packages that are installed on the system 16:43:19 Hi, can I compile more than one port at a time or should I have one compile after the other? 16:49:07 the easy way is to use poudriere to build a package repo as it takes care which ports can be build in parallel and other stuff 16:53:49 Anyone use CBSD? I'm following the guide for a bhyve guest, and as soon as I get to "cbsd login my_guest" it returns error message "no such CBSDMake file: /root/CBSDfile" 17:00:16 I do see that there is a "CBSDfile" located at /usr/local/cbsd/share/examples/cbsdfile/CBSDfile, but not mentioned in the docs at all 17:01:43 ravella: I do 17:01:54 the command is not login, but blogin for bhyve and jlogin for jails 17:03:01 oh dear, I probably need new glasses, the docs do indeed say "blogin" but I read it as "login" 17:03:03 login for, as you discovered, CBSDfile, which is something like Vagrantfile or docker-compose.yaml, if you're familiar with those technologies (if not, just ignore I said that) 17:03:12 thanks meka! 17:03:14 *login is 17:03:19 ravella: you're welcome! 17:27:18 i'm having some trouble with bhyve and vlan interfaces i'm using vm-bhyve fwiw but i'm not sure it's related to it. basically if i have a bridge with members bge0.5 and i configure the vm as a member of that bridge, i can reach the host VLAN ip but not other hosts 17:27:33 but from other hosts i can reach the hypervisor host just fine 17:27:52 am i missing some sysctl tunable related to allowing bridges to pass traffic? 18:01:26 is it up? ^^ 18:01:54 hjf, forwarding ? 18:02:08 net.inet.ip.forwarding=1, ^ 18:02:20 yeah 18:08:26 cpet: i tried that but it didn't work. i don't think it's related though because bridges are at a lower layer than ip fowarding 18:25:31 hjf: can you share `vm switch list` ? 18:26:29 NAME TYPE IFACE ADDRESS PRIVATE MTU VLAN PORTS 18:26:30 public standard vm-public - no - - bge0 18:26:30 b3k standard vm-b3k - no - 3000 bge0 18:27:28 NAME TYPE IFACE ADDRESS PRIVATE MTU VLAN PORTS 18:27:29 sw100 standard vm-sw100 - no - - lagg0.100 18:27:54 I just created the switch and added my nic to it 18:27:54 hrm 18:28:28 i created the vlan on top of the nic and then added _that_ to the switch 18:29:05 yeah this isnt that fancy vlans are done using the actual switches webui 18:29:05 which was working for me before i changed the network around and bridged everything. 18:29:35 well you want it tagged properly bfore it leaves bge0 right? 18:30:25 hjf, is the switch name the same as what you created ? 18:30:32 in the .conf ? 18:30:36 yes 18:30:57 and its basically vm switch create public && vm add public nic ? 18:31:09 $public = whatever 18:31:28 well i edited the conf 18:31:58 what you create is what you add cant add public and add in slappy to the conf 18:32:36 https://www.gsp.com/cgi-bin/man.cgi?topic=VM-BHYVE 18:32:37 Title: VM-BHYVE(8) - utility to manage bhyve virtual machines 18:32:39 Quick Start 18:32:43 switch_list="sw100 sw200 sw30 mgmt host" 18:32:43 ports_sw100="lagg0.100" 18:32:44 is all you need to do 18:33:27 ^_+ 18:33:28 lo 18:33:34 once that works you can go all crazy with vlans and all that 18:33:59 network0_switch="public" 18:34:03 network1_switch="b3k" 18:34:16 so yes the vm has both "switches" 18:34:50 still have to configure the OS you install 18:35:06 the os has ip addresses for both interfaces 18:35:11 oh i didn't ask that; do you want the host or the vm to do the tagging 18:35:20 the host 18:35:36 then i think you have to add the vlan to the switch 18:35:37 the linux if attached to public gets an ip from dhcp fine 18:36:06 the if attached to b3k has a static ip. it can ping the hypervisor host (in this case the ip i set for bge0.3000) 18:36:24 but it can't ping hosts behind bge0.3000 18:36:49 hosts connected to the switch can ping bge0.3000 , but not the VMs 18:43:49 i changed it a little so it looks like this: 18:43:50 NAME TYPE IFACE ADDRESS PRIVATE MTU VLAN PORTS 18:43:50 public standard vm-public - no - - bge0 18:43:50 b3k standard vm-b3k - no - - vlan3000 18:44:20 but the result is the same, i can ping the hypervisor host but not the remote hosts (and the remote hosts can ping the hypervisor but not the VMs) 20:50:07 hjf: i feel like i have been here before; there were three or four variants of configuring vlans and the vlanXX wasn't it 21:56:45 meka: do you use a rfc1918 network for your guests? The nat situation with cbsd is giving me some trouble 21:59:47 I'm considering manually creating the bridge instead, and adding a NAT rule to pf 22:02:35 anyone run wget in a cron job? do you have to run it so there is no output printed? 22:04:59 the command runs normally but did not in the cron, so it seems like that cron likes commands that don't produce any output 22:05:18 wget -q option 22:10:50 or do i need to add a new line after the command? 22:26:32 ravella: that's exactly what I did with Reggae 22:27:34 To be precise, there's one special jail and what makes it special is dhcp/dns inside of it. Other jails get their addresses through dhcpcd, as it picks up IPv4 and IPv6 22:28:32 rtprio: i've been using ifname_vlans="vlanNNNN" attached to a virtualbox VM for years 22:28:38 no switch though 22:28:50 On the host I have bridge0 renamed to cbsd0 just to make it easier to identify the bridge which is for CBSD jails and I have all my jails as VNET, so when I create the jail with jcreate, I add interface=cbsd0 22:30:00 An example of such jail creation is this line: https://github.com/cbsd/reggae/blob/master/scripts/master-init.sh#L26 22:30:01 Title: reggae/master-init.sh at master · cbsd/reggae · GitHub 22:30:06 i had to tcpdump every interface until i could figure out whr th problem was 22:33:27 "oh noes" 23:09:43 so now i changed things a little. i assigned the host ip to the bridge, not the vlan interface 23:10:01 bridge includes vlanif + tap 23:10:12 if i do it this way, i can't ping from outside 23:10:55 the vm and the host see each other ,but not through the (physical)switch 23:14:45 ok it gets worse. i flushed the arp cache, and i tried pinging. the destination MAC appeared in the arp cache 23:20:25 meka: ah I think I found my problem, if I create the bridge manually then I need to answer "no" to `cbsd natcfg` 23:25:09 SON OF A B 23:25:43 When you use 802.1Q VLAN tagged interfaces created on a physical interface (parent interface), make sure not to add the parent interface to any bridge. 23:25:46 If you do this, ARP resolutions on the tagged interfaces seem to cease working as reported in the following PR/Forum Threads. 23:25:58 indeed, i did just that and now it works 23:31:28 hjf: wouldn't surprise me if you and I just had similar issues lol 23:32:06 well, in my case, I couldn't ping the outside world, two daemons were contending to create the same bridge interface 23:32:24 yeah, on second though, probably not the same issue 23:32:58 this sucks though because i need the parent interface in a bridge 23:33:27 maybe a nasty solution would be to make another tagged vlan interface for the default vlan (vlan 1?) 23:34:57 so all my traffic will be VLANed. i guess it's ok as long as i always have a switch that can deal with this