00:49:34 * AmyMalik mumbles something about PCI
05:01:46 <sfox> Hello I need help
05:02:15 <sfox> my jail host is somehow allowing jails which are assigned their own interface parented to a vlan to access management vlan hosts
05:02:55 <sfox> it's bypassing the router and it looks like the freebsd jail hypervisor itself is forwarding packets to the management vlan, despite never putting gateway_enable="YES" into rc.conf
05:05:25 <sfox> also, traceroute is contradicting ifconfig saying an ip address is not on a interface
05:06:57 <sfox> net.inet.ip.forwarding: 0
06:26:38 <meena> sfox: what's your setup look like?
06:29:49 <sfox> meena, what do you mean exactly?
06:29:53 <sfox> it's a server
06:30:12 <sfox> sorry, confused
06:30:20 <sfox> you want an ifconfig report or something?
06:46:07 <meena> sfox: what does your jail setup look like? are those vnet jails? what does your topology look like… etc
06:46:34 <sfox> i'm using standard iocage stands. vnet is off
06:47:10 <sfox> my topology is the main hypervisor operating system's ipas on at a lagg0 interface in failover mode across to intel gigabit nics pointed to vlan1
06:47:44 <sfox> the iocage jails are assigned to an alias interface 'jail' each jail with a unique ipv4 and ipv6 address
06:48:05 <sfox> the jail interface is renamed lagg0_4 or vlan4 on interface lagg0
06:48:42 <sfox> all those cluster physical links go into a layer 3 switch, which is trunked to a openwrt router which handles firewalling and cross-vlan talk, nat, etc
07:04:39 <sfox> meena, make sense?
07:04:48 <sfox> *standard iocage jails.
07:57:53 <meena> sfox: so far so good. Still not entirely sure what *exactly* is happening that should not be happening
07:58:24 <sfox> thankyou
08:01:52 <meena> what kinds of connections are the jails making that your setup and topology disallows?
08:09:33 <sfox> my computer which is on the management vlan
08:09:38 <sfox> 192.168.55.5
08:10:03 <sfox> i ran nc -vkl 4000 which binds to 0.0.0.0 port 4000 TCP
08:10:39 <sfox> and inside one  of the tails (192.168.4.31) i ran nc -v 192.168.55.5 4000
08:10:59 <sfox> the connection succeeded and on my computer I see these lines:
08:11:08 <sfox> Listening on 0.0.0.0 4000
08:11:08 <sfox> Connection received on 192.168.4.31 12606
08:11:08 <sfox> hello
08:11:15 <sfox> that should not happen
08:11:48 <sfox> i thought my router was passing it through so I connected my laptop to the switch vlan 4, and tried to repeat nc -v 192.168.55.5 4000
08:11:55 <sfox> i was unable to replicate
08:12:33 <sfox> i then on the jail hypervisor ran traceroute -i jail -s 192.168.4.31 192.168.55.5
08:13:09 <sfox> traceroute complained the jail interface didn't contain that IP address, which directly contradicts what ifconfig jail says
08:13:23 <sfox> i removed -i jail and ran traceroute again and it succeeded
08:13:30 <sfox> that shouldn't happen either
08:13:40 <sfox> interestingly there was no router inbetween
08:14:09 <sfox> according to traceroute, it was direct. That lead me to believe the jail hypervisor is forwarding the packets
08:14:28 <sfox> which also should not be possible because sys.inet.forward=0
08:14:43 <sfox> sys.inet.forwarding is 0
10:37:26 <Grabunhold_> i'm trying to install py39-psycopg in order to manage my postgresql with ansible. however, it seems py39-psycopg is built against postgresql13-client.
10:37:26 <Grabunhold_> so when trying to install py39-psycopg, it forces removal of postgresql15-server because that depends on postgresql15-client which in turn conflicts with postgresql13-client. is there a way to get both postgres15 and py39-psycopg?
10:38:52 <parv> Compile the ports & respective dependencies yourself
10:42:24 <parv> ... instead of using FreeBSD-provided packages
11:00:23 <Grabunhold_> well, pkg is really easy to use from ansible. i'll have to look into how to build the port from ansible.
11:05:46 <Grabunhold_> there is https://docs.ansible.com/ansible/latest/collections/community/general/portinstall_module.html but how do i set compile-time options...
11:05:47 <VimDiesel> Title: community.general.portinstall module – Installing packages from FreeBSD’s ports system — Ansible Documentation
11:06:44 <Grabunhold_> building the port has the pleasant side-effect of disabling the dependency on the C client libs altogether and roll a pure python implementation, performance is of no concern anyway
11:25:31 <mage> Grabunhold_: I suggest to setup your own poudriere repository
11:27:27 <Grabunhold_> just to create a pgsql database from ansible? oof
12:51:17 <sopparus> hello, anyone using rsnapshot here? is there a sane way to backup a backup point?
12:51:30 <sopparus> as in duplicate your backup on another server or another hdd
13:01:26 <Grabunhold_> sopparus: not using rsnapshot here, but you might want to look into zfs send / recv
13:22:38 <meena> sfox: vlan is a link layer thing, tho, so sys.inet.forward=0 should have almost nothing to do with it
13:23:15 * meena might be talking out of her ass tho… 
13:26:44 <sopparus> yeah, i'll do that Grabunhold_. thanks
14:04:54 <wikan> hi. I noticed network issues on freebsd
14:05:09 <wikan> when I download something - for example file 1, file 2, file 3
14:05:20 <wikan> i noticed the first file has almost 100% of bandwith
14:05:55 <wikan> another thing is sshfs. When I send files to remote machines, Firefox can;t connect websites
14:06:07 <wikan> should I setup something?
14:41:53 <CrtxReavr> wikan, are you on say. . .. a 10 Gbit network?
15:47:38 <wikan> CrtxReavr i am sorry i was sleeping
15:48:20 <wikan> Via LAN it is maximum network card transwer, about 10MB/s
15:48:41 <wikan> via WAN obout 6MB/s
15:56:17 <gzar> hi, im getting kernel crashes/panics when trying to open an SDL window
16:03:38 <gzar> pretty sure i missconfigured something when installing the nvidia driver
16:12:30 <CrtxReavr> You're using the binary blob x11/nvidia-driver, or something else?
16:38:17 <gzar> binary blob
16:38:49 <gzar> i recompiled sdl2 without vulkan, maybe that was causing it. once i turn on kernel crash dumps i'll try again
16:42:06 <mvanbaak> hey all. What's the big difference between '/usr/local/rc.d/<somethnig> start' and 'service <something> start' ?
16:46:59 <mvanbaak> https://github.com/morpheus65535/bazarr/issues/2104
16:47:01 <VimDiesel> Title: TrueNAS: Does not autostart · Issue #2104 · morpheus65535/bazarr · GitHub
16:47:01 <VimDiesel> 2104 – man pages for modunload(8) and modstat(8) have incorrect references https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=2104
16:47:21 <mvanbaak> this person states that `/usr/local/etc/rc.d/bazarr start` works, and `service bazarr start` doesn't
16:57:07 <CrtxReavr> mvanbaak, I would suspect the messed with local_startup in rc.conf
16:57:12 <CrtxReavr> they messed
16:58:31 <CrtxReavr> or /usr/local/etc/rc.d/bazarr itself.
17:01:31 <yuripv> mvanbaak: check ENVIRONMENT section in service man page
17:11:37 <Beladona> Hi
17:12:30 <wikan> let me ask my question again
17:12:44 <wikan> i notices network issues under freebsd
17:13:35 <wikan> for first, when i was uploading files to remote machine via sshfs, Firefox couldn't load any website. Some of them displayed "You are in offline mode" message.
17:14:06 <mvanbaak> yuripv: gheh, and with '/usr/local/etc/rc.d/<service> start' this is not the case? thought they were the same
17:14:12 <wikan> second, when I download more than one file, I noticed first one use 95% of bandwith
17:14:19 <mvanbaak> so I need to set PATH somehow for the bazarr port
17:16:56 <wikan> should I configure something?
17:46:26 <asoija> hello, I'm trying out qemu to run a few linux boxes from 13.1-RELEASE. Everything seems to go OK except for one thing: the VMs don't have network connection. I tried both using the default slirp adapter and setting net.link.tap.user_open=1 as per the pkg message for qemu, but either way no carrier media is detected on the guest
17:47:46 <asoija> This is how I'm starting the VM (as simple as possible for the network test) -> https://termbin.com/gl07
17:49:23 <asoija> I read several posts in the forums covering differing aspects of qemu virtualization but to no avail. So I guess the question is: what's the simplest way to set up networking in qemu guest VM from a freebsd host?
17:51:55 <Beladona> both mirror drives have: 1  efi  (260M) 2:  freebsd-boot  (512K)   under `gpart show`. One boots, other does not. Why?
18:10:33 <rukus> boot flag
18:10:54 <Beladona> rukus what?
18:16:08 <CrtxReavr>  /usr/ports/.portsnap.INDEX is all just sha256 sums?
18:19:50 <Beladona> rukus  status` says `gpt/efiboot0     N/A  da1p1`  but no `efboot` for da2
18:20:40 <BaloneyGeek> Is the pkg git package known to be broken right now? I just tried to clone a GitHub repo over HTTPS and got `git: 'remote-https' is not a git command. See 'git --help'.`
18:29:21 <mvanbaak> hum, I really dont get it
18:29:50 <mvanbaak> for some reason for me, in a completely new jail, something that is started using the `service` command can detect binaries in /usr/loca/bin
18:30:34 <mvanbaak> as if somethintg on my machine sets the PATH
18:30:37 <mvanbaak> but I cant find where
18:34:45 <yuripv> mvanbaak: probably it's using the absolute path?
18:37:27 <mvanbaak> nope. it's really weird
18:38:03 <mvanbaak> I created a port of this python tool. it does not have the paths hardcoded. the python code uses which to find out where unrar is installed. which is in /usr/local/bin
18:38:40 <mvanbaak> according to the service(8) manpage, the path does not contain this. which also explains why this truenas user stats it is not working for them with the service command
18:38:49 <mvanbaak> but I try to reproduce it, and on my machine ItJustWorks
18:38:51 <mvanbaak> lol
18:42:04 <yuripv> that's all what service really does (it's a shell script): exec env -i -L -/daemon HOME=/ PATH=/sbin:/bin:/usr/sbin:/usr/bin "$dir/$script" "$@"
18:42:30 <rtprio> mvanbaak: are you porting unpackerr?
18:45:32 <yuripv> you know what's weird, it uses the daemon login class, and it seems to override the PATH set, i.e. try running `env -i -L -/daemon HOME=/ PATH=/sbin:/bin:/usr/sbin:/usr/bin echo $PATH`
18:49:38 <yuripv> nevermind me, bad usage :D
18:51:23 <mvanbaak> rtprio: no, trying to find out why bazarr port will not start on a 13.0 jail on truenas, but it does work in my jail on plain FreeBSD
18:52:07 <mvanbaak> yuripv: in my /etc/login.conf I do have /usr/local/bin in the path of the default class (and the daemon class as that one gets it from the default class
18:52:26 <mvanbaak> as if on my machine 'service' leaves the path alone
18:55:13 <yuripv> i tried running a script using that env invocation that prints HOME and PATH into a file, and that file confirms what man page says
18:56:23 <yuripv> i.e. https://pastebin.com/MHbizYBa
18:56:25 <VimDiesel> Title: $ cat test.sh#!/bin/shecho $HOME > 1echo $PATH >> 1$ exec env -i -L -/ - Pastebin.com
19:03:01 <rtprio> ah; that's an *rr i wasn't aware of
19:05:40 <mvanbaak> /sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin://bin
19:05:44 <mvanbaak> that's the path I have
19:05:50 <mvanbaak> not in your test script
19:06:13 <mvanbaak> but if I replace the bazarr.py in the bazarr package with https://pastebin.com/hJppvfrC
19:06:14 <VimDiesel> Title: # scriptwith open('/tmp/bazarr.txt', 'w') as fp: fp.write(os.environ.get( - Pastebin.com
19:08:46 <mvanbaak> this looks a lot like the default class in login.conf:         :path=/sbin /bin /usr/sbin /usr/bin /usr/local/sbin /usr/local/bin ~/bin:\
19:09:10 <mvanbaak> does the daemon command do maybe?
19:09:46 <yuripv> are you running it using daemon?
19:10:13 <yuripv> i mean inside your rc script
19:10:55 <parv> "//bin" -- some variable is not set?
19:12:47 <mvanbaak> yes, it's running via daemon in the rc script
19:14:01 <mvanbaak> https://cgit.freebsd.org/ports/tree/net-p2p/bazarr/files/bazarr.in
19:14:03 <VimDiesel> Title: bazarr.in « files « bazarr « net-p2p - ports - FreeBSD ports tree
19:16:03 <Beladona> both mirror drives have: 1  efi  (260M) 2:  freebsd-boot  (512K)   under `gpart show`. One boots, other does not. `glabel status` says `gpt/efiboot0  da1p1`  but no `efboot` for da2.  Why? and how to make that bootable too?
19:25:46 <mvanbaak> I am totally lost
19:26:18 <debdrup> same.
19:37:45 <Beladona> Do I need to do `gpart bootcode -b /boot/bootxyz da2` ?
19:38:46 <mvanbaak> yuripv: os.environ in python has this path when I use 'service bazarr start'
19:39:32 <mvanbaak> I replaced the bazarr.py that is called a python script that does write os.environ to a file in tmp. It does not load any of the bazarr files just to make sure.
19:39:34 <mvanbaak> environ({'VENDOR': 'amd', 'LOGNAME': 'bazarr', 'PAGER': 'less', 'LANG': 'C.UTF-8', 'OSTYPE': 'FreeBSD', 'MACHTYPE': 'x86_64', 'PATH': '/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin://bin', 'EDITOR': 'vi', 'HOST': 'bazarr-test', 'PWD': '/', 'GROUP': 'bazarr', 'USER': 'bazarr', 'HOME': '/', 'HOSTTYPE': 'FreeBSD', 'MM_CHARSET': 'UTF-8', 'BLOCKSIZE': 'K', 'RC_PID': '32770', 'SHLVL':
19:39:36 <mvanbaak> '1'})root@bazarr-test:~ #
19:39:51 <Beladona> ref https://docs.freebsd.org/en/books/handbook/geom/#geom-mirror-two-new-disks
19:39:52 <VimDiesel> Title: Chapter 20. GEOM: Modular Disk Transformation Framework | FreeBSD Documentation Portal
19:40:20 <mvanbaak> I'm totally out of ideas here
19:52:05 <mvanbaak> just to make sure I simply restarted the jail, but it has the exact same info in the environment
20:03:57 <Lovis_IX> mvanbaak:you said that your try your script on a 13.x jail on truenas and it does not work. But it works on your dev machine which run 13
20:04:04 <Lovis_IX> .x
20:05:39 <Lovis_IX> [sorry]. Trunas is still in 12.x iirc. ans a host could not have a have with an os version superior has his one no ? Or I am adding confusion on confusion?
20:09:03 <ek> TrueNAS is on 13.x these days.
20:09:44 <ek> 13.1-RELEASE, I believe.
20:10:49 <Lovis_IX> Oh!…. Need to make some upgrade to my NAS. Thanks and sorry for the bad noise.
20:11:59 <ek> Lovis_IX: Not a problem. Glad you can get some updates going. :)
20:18:23 <mvanbaak> gheh, no problem
20:18:50 <mvanbaak> I did not test with truenas, but that's what the user reported
20:18:56 <mvanbaak> they said they run 13.0 in the jail
20:19:25 <Lovis_IX> ok
20:20:51 <mvanbaak> I'm the bazarr port maintainer, so the bazarr dev asked me to look at it :) but I cant reproduce it on freebsd. So I'm trying to find out where this PATH is set in my freebsd setup, as it does NOT match with either /etc/login.conf classes nor with the manual page of service(8)
20:21:59 <mvanbaak> LOL, they found out how to fix it on their end
20:22:35 <mvanbaak> Finally figured it out.. By pure habit I change my shell to zsh upon creating a jail and fixing som bindkey issues I have with my terminal
20:22:37 <mvanbaak> Setting the terminal back to csh fixed the issue. Sorry for not picking up on that detail untill now..
20:22:50 <mvanbaak> but this doesn't change the fact I cannot find where the paths are set for me
20:22:56 <mvanbaak> but at least I can go to bed now
20:22:58 <mvanbaak> hahaha
20:31:21 <Lovis_IX> mvanbaak: have a good night, whithout nightmares.
21:09:00 <RoyalYork> Do you guys primarily use a desktop enviornment or are you running solely through the terminal?
21:09:22 <vkarlsen> I primarily use fluxbox
21:10:30 <RoyalYork> I got the desktop working on my end (xfce). Was unsure if most people are using a DE or solely the terminal
21:11:08 <la_mettrie> most of people probably browse the web
21:11:51 <RoyalYork> Im running 13.1 in a virtual machine testing it as a daily driver.
21:12:17 <RoyalYork> Trying to see how much I am missing from windows and so far it doesn't seem like im missing all that much
21:12:20 <RoyalYork> youtube works
21:12:22 <RoyalYork> netflix works
21:12:33 <RoyalYork> officelibre works
21:12:39 <RoyalYork> firefox works
21:12:42 <RoyalYork> etc
21:27:40 <wwt> firefox with sound even? ;)
21:31:18 <mason> RoyalYork: Eh? How did you get widevine running?
21:32:24 <RoyalYork> wwt, yes firefox with sound. widevine? no idea. LoL
21:32:28 <RoyalYork> I just followed the handbook
21:33:07 <RoyalYork> I may be lucky because im using VMware video drivers and "intel audio" through virtual box
21:33:20 <RoyalYork> but I didn't do anything special
21:36:21 <wwt> RoyalYork: the freebsd handbook?
21:36:31 <RoyalYork> correct
21:36:55 <wwt> netflix via linuxulator?
21:37:37 <RoyalYork> No. Nothing like that. I installed 13.1 on virtualbox within a windows 10 host. Installed xfce and it just worked
21:38:03 <RoyalYork> You got me thinking, maybe I should start another box and attempt to replicate this
21:44:54 <RoyalYork> hold on
21:44:58 <RoyalYork> just tested it
21:45:08 <RoyalYork> i only got to the movie menu in netflix
21:45:16 <RoyalYork> doesn't play scrap that
21:45:23 <RoyalYork> I was confusing the video with my jellyfin player
21:45:39 <RoyalYork> user error!!!
21:47:49 <mason> RoyalYork: kk, no worries, I was curious. What's missing there that's preventing your netflix from working is widevine.
21:53:59 <jarebear6expepjo> welcome!
22:19:10 <RoyalYork> mason, , thanks ill look into widewine
22:19:30 <jarebear6expepjo> or consider whether you want to support it, too
22:19:40 <jarebear6expepjo> drm bad btw
22:22:55 <mason> RoyalYork: widevine, and the most common way I've seen people approach it on FreeBSD is to run Google Chrome in the Linuxulator, with varying degrees of completeness (audio) and success.