[05:36] <Bebef> Okay, so I think I have tried everything, but I can't get the Ubuntu installer to work.
[05:37] <Bebef> I see the boot menu, then the kernel messages, but when the installer itself starts, I end up with a black screen. Same happens on two different displays. Signal is active (the displays aren't complaining "no signal") but the screen is black and I can't do anything.
[05:38] <Bebef> I reckon it's some issue with the framebuffer console.
[05:39] <Bebef> Is there any way to influence/change the installer to a different graphics mode?
[05:40] <lotuspsychje> wich ubuntu release are you trying on what kind of machine/server ?
[05:42] <Bebef> It's 22.04 server on a HP MicroServer Gen10. The desktop installer works, funnily enough...
[05:42] <lotuspsychje> hmm
[05:43] <Bebef> Just firing up the 20.04 installer, just to see if that behaves differently in any way.
[05:43] <Bebef> I assumu that, when I can run the desktop graphical installer, it shouldn't be an issue regarding the grapics card settings, like VRAM.
[05:44] <lotuspsychje> its weird, as subiquity should rather work better then a GUI installer on a server
[05:44] <Bebef> Yes. 20.04 is the same btw.
[05:44] <lotuspsychje> can you F1 at setup boot to see at wich point it bottlenecks?
[05:45] <Bebef> I start to wonder how I installed the machine in the first place 
[05:45] <Bebef> lotuspsychje: Start the installer via grub menu, then hit f1?
[05:46] <lotuspsychje> once the installer start try to hit F1
[05:46] <lotuspsychje> maybe we can catch text boot errors
[05:47] <Bebef> hmm. I briefly see the kernel messages and the screen turns black after that
[05:47] <lotuspsychje> whats the last line you see
[05:47] <lotuspsychje> maybe catch it in a real screenshot and share to the volunteers here
[05:48] <Bebef> Can't really say, it's so quick
[05:48] <Bebef> I wanted to try the serial console yesterday, but the PCIe card didn't survive shipping :(
[05:48] <lotuspsychje> there's no uefi settings in that?
[05:49] <lotuspsychje> or other bios settings that could prevent the installer going through?
[05:49] <lotuspsychje> usualy the ubuntu installer doesnt stall at the very start
[05:50] <Bebef> I wouldn't know what setting that would be. I mean, if it was BIOS/UEFI, the kernel wouldn't boot, right?
[05:50] <Bebef> I assume it's an issue with my graphics hardware and a mode that's chosen, but not supported. Something the like.
[05:50] <lotuspsychje> well black screens at boot level sounds uefi
[05:50] <lotuspsychje> !nomodeset
[05:50] <Bebef> Not at boot. The Installer!
[05:50] <lotuspsychje> can try this too^^
[05:51] <Bebef> Like I said, I can boot the USB and I can see the kernel messages. Once the installer itself starts, the screen turns black.
[05:51] <lotuspsychje> weird
[05:52] <lotuspsychje> have to bbl a bit, try to catch errors in a screenshot or !paste Bebef 
[05:53] <Bebef> Yeah, me too. Just testing the 18.04 installer, which doesn't work either...
[05:55] <Bebef> hah yes!
[05:56] <Bebef> Thanks to the nomodeset, I can now see the systemd boot output
[05:56] <Bebef> And I can see the installer!
[06:27] <Bebef> Oh boy, now that I have a running installer, it crashes after partitioning :(
[06:50] <Bebef> Is there a (easy) way to modify the Ubuntu USB installer so it already includes "nomodeset" in the boot parameters?
[07:17] <lotuspsychje> !yay Bebef 
[07:17] <lotuspsychje> !yay
[07:30] <lotuspsychje> Bebef: try to catch subiquity in text mode to see why it crashes under the hood
[07:31] <Bebef> Yeah, probably my own fault, tried to use pre-configured (partitioned) disks. I discarded everything and installed everything from scratch with blank disks, now it works.
[07:31] <nirox> Ubuntu 22.04 netplan question. I have two interfaces, ens192 with public IP and ens224 with 10.10.0.2 IP. im using "routes to 10.10.0.0/24 on-link: true" and also tried "scope: link" on ens224, but it does not add this route and im unable to ping two machines with same config. But if i do "ip route add 10.10.0.0/24 dev ens224" i get correct route and stuff works. I do not have a GW/router 
[07:31] <nirox> on 10.10.0.0/24.
[07:33] <frickler> nirox: what netmask do you have on ens224? do you use dhcp or are these static addresses?
[07:36] <nirox> Static IP, "- 10.10.0.2/24"
[07:43] <nirox> https://pastebin.com/TTw0n2bB - This is netplan server1, tried to reorder ens224 before ens192 just to make sure it wasnt order that mather.. server2 is same, but ens224 10.10.71.12 and ens192 123.123.123.102.. This works after doing "ip route add 10.10.71.0/24 dev ens224" that i understand should be equal to "scope: link"/"on-link: true"
[07:43] <weedmic> nirox: every/each subnet needs its own gateway.  the one with the public IP you need to get from them.  are you sure you have the right gateway?
[07:44] <nirox> Public is working correct
[07:44] <nirox> Its the LAN 10.10.71.11<->10.10.71.12 i do not get to work before i do "ip route add 10.10.71.0/24 dev ens224"
[07:45] <weedmic> ok, that's the hard one - the other is what you set.  is it a vm or physical device?
[07:45] <nirox> VM
[07:45] <weedmic> nat or bridged?
[07:46] <nirox> Same physical virtual host. Just virtual switch
[07:46] <weedmic> seems to me you should pick bridged - in the host esxi (or equivalent) you need to assign a bridged adapter to the vm.  inside the vm, just ensure it is set to dhcp.  it will work (if the host has connectivity which it does).
[07:48] <weedmic> if you decide to use NAT, you need to create the NAT device on the esxi (or equivalent) which many people forget - but bridged is really best unless you have many machines and want to assign IPs manually etc.
[07:48] <nirox> Hum, as i see it the problem is that ubuntu never set which interface it should use and it uses default that is public ip net..
[07:49] <nirox> because when i manually set 10.10.71.0/24 should use ens224, it works. So how can changing on VM attached interface change this behave?
[07:49] <nirox> I do want both the public IP and the private IP. Because there is internal cluster data in applications ont this servers that should talk via LAN
[07:52] <weedmic> nirox: for the most part, a vm is the same as a physical machine.  but... you did not actually plug the vm into anything.  so you have the tell the "chassis" what it is plugged into.  what are you using?  virtualbox vmware?
[07:52] <weedmic> I can walk you through
[07:52] <nirox> ESXI
[07:52] <weedmic> ok, let me open that
[07:55] <nirox> I need to ask again if you got the part that when i do "ip route add 10.10.71.0/24 dev ens224" everything is working as i want. its just that netplan does not apply that..
[07:55] <weedmic> under virtual machines, highlight the machine you need to edit, click edit, go down to "network adapter"> click LAN (you might have WAN), check connect at poweron - let me know if it was wan
[07:55] <frickler> 10.10.0.0/24 and 10.10.71.0/24 are different subnets
[07:55] <weedmic> everysubnet needs a gateway
[07:55] <nirox> yea, forget 10.10.0.0/24. it was just a lacy example. im now using the correct range
[07:56] <weedmic> gateway is not a mask /24 is a mask
[07:56] <nirox> Sorry if that made it confused. i posted the correct netplan file as link
[07:56] <nirox> No, you should not need a GW if its a LAN that talks on same mask
[07:57] <weedmic> np, at all.  correct, but.. you have two lans
[07:57] <frickler> not sure why you need a routes: statement at all. the normal onlink connected route will be added automatically like for any network
[07:58] <nirox> I tried without defining routes, but it does not work before adding it manually..
[07:58] <weedmic> I have now checked a few machines, I do not have such in my ip tables either
[07:58] <weedmic> just sets in esxi and inside the machine.  
[07:59] <weedmic> but, seriously, ensure the esxi chassis for networks is correct - else you are wasting your time.  
[08:00] <weedmic> if it is, then remove that line in iptables and ensure each subnet has a gateway (and mask - which i'm sure you have).
[08:03] <weedmic> u can use this command (change the nic) to see the items grouped nicely - ip -4 a show enp0s31f6
[08:05] <nirox> There is no "LAN"/"WAN" in network settings. Everything is by esxi a "LAN" afaik, as a public IP is also a LAN untill its routet into the internet
[08:06] <weedmic> I have both - interesting - however, you have ensured it is LAN, so the problem is inside the VM.
[08:06] <nirox> In esxi its a VLAN on the "Distributed virtual group"/"DSwitch"
[08:07] <weedmic> from esxi interface, does it say "connected = yes"?  also what does it say under "Pass-through (Direct-path I/O)"?
[08:08] <nirox> yes, both enabled
[08:09] <nirox> Using vcenter, but also checked directly on host - there is nothing about LAN/WAN
[08:11] <weedmic> do you have other vms that work on that network?  
[08:12] <nirox> https://postimg.cc/ts7RDW0b - this is dswitch layout.
[08:12] <weedmic> esxi>network>LAN>edit settings>virtual switch - what do you see?
[08:13] <weedmic> oic, under the physical adapters, did you setup those under their tabs after creation?  they need to have the correct range entered.
[08:14] <nirox> 10.10.71.0/24 on physical adapters?
[08:17] <weedmic> you should be able to click on the one you need to ensure is correct, for example vmnic2 and it should open the sets - or you can go to the tab and find it and open the sets.  down near the bottom (a lot of scrolling) it should say the subnet data (probably called "discovered subnets".  is anything there?
[08:17] <nirox> No, its not listed in "discovered subnets"
[08:18] <weedmic> this might be your prolbem.  for me, this is what I see (many of them - i'm just going to pick the one I'm on right now)
[08:19] <weedmic> i click on my machine in the list, go to network, click on LAN, it opens up the LAN that it is on - i see a graphical representation of a switch, 5 ports are empty, but the others are full, i click on the "physical adapters" and see how that switch was seutp.  at the bootm I see my seubnet.
[08:21] <weedmic> I should guess that if you do not see a lot of things there, you need to setup a LAN.  but.. I should not delete the one you have now.  Just create a new one called perhaps LAN0, setup the LAN with your sets and then connect the VM to that LAN0.
[08:21] <weedmic> or reask the question and someone could offere a different path to solution.
[08:22] <nirox> We have the same on other VLAN. where we just use LAN IP (10.10.53.2+3), but have GW. trace goes from 10.10.53.2 to 10.10.53.3 directly, not via GW. But route shows up for dest 10.10.53.0 on iface
[08:23] <nirox> The problem i see, is that netplan does not set route for dest 10.10.71.0, it uses default route and thats going thrue wrong interface on VM guest
[08:24] <Bebef> Next issue. I have full disk encryption running, so I set up dropbear-initramfs and configured everything. In the boot screen, I can see "Begin: Starting dropbear ... " right after the NIC config. I can ping the right IP, but there seems to be no SSH running.
[08:25] <nirox> If i set route via ens224 manually, it works. so i do not think its the vm host the prblem.. i will check more and give next issue your time. thanks for help
[08:26] <weedmic> np, i'm here to help - but I concur it is not the host if you can manually get it to work inside the vm - i just have a set way to hone things down - always starts with hardware (which in a vm is the esxi).
[09:29] <nirox> My solution will be to use ifup-hooks with "ip route add 10.10.71.0/24 dev ens224". Allready doing this for another problem where i do not want autoadded route for public ip to work without going via GW, as machines within same /26 is isolated to talk to each other directly - only possible via GW/FW rules..
[10:31] <nirox> Uh, it was this fix that without me knowing it beeing added to this servers too :facepalm: Script that deletes public route was not specific enough, so it delete all routes except default.. :some-embarrassed-emoji:
[11:18] <faekjarz> Gidday, mates! Anyone competent in virtio networking? I'd like to control the throughput of a virtual network IF, to numbers below 100/fast ethernet. It's software defined networking, right? …i'd like to simulate a "variable width pipe", down to 1mbit/s. Which search terms would you suggest?
[11:20] <faekjarz> i hope this is feasible on the libvirt level - if not, well, i'll set up another OPNsense/OpenWRT VM, which does the throtteling
[11:26] <faekjarz> I've x-posted in the #virt channel …yeah, that's more on-topic
[11:27] <weedmic> faekjarz: define "virtual network IF"
[11:28] <weedmic> and are you using vmware or virtualbox?
[11:29] <faekjarz> weedmic: i'm invoking the virtualisation context, libvirt / virt-manager - THOSE virtual network interfaces
[11:29] <weedmic> ok if=interface
[11:29] <faekjarz> aye
[11:29] <weedmic> and are you using vmware or virtualbox?
[11:30] <faekjarz> weedmic: neither, it's vanilla libvirt / virt-manager …some call it Linux-KVM
[11:30] <faekjarz> (qemu)
[11:34] <weedmic> i have aqemu - hoping it is gelijk
[11:35] <weedmic> alike
[11:38] <weedmic> hmm - i regret to inform you I saw no options like what request - actually, I see it is under the phone project, so it was part of the plan to make a vm on a phone.  others have read it, give them time to help.
[11:40] <faekjarz> weedmic: danpb in the #virt channel just posted a link - https://libvirt.org/formatdomain.html#quality-of-service - this seems to do exactly what i'm looking for.
[11:51] <faekjarz> thanks for bothering, anyway, weedmic :) 
[12:01] <weedmic> tks for reporting back - very nice
[12:10] <faekjarz> sure, this is open source! <3 …everyone should help, proliferating education ("back-porting" relevant posts from other channels) …almost, but not quite, like Leninism - everyone should know how the "beans of production" work xD LOL
[15:31] <J_Darnley> Is there a way to specify "Before" in systemd "A.service" all instances of another "B@.service"?
[15:31] <J_Darnley> Can I use a glob in these files like I can on the command line?
[16:51] <Niclos98> Hello  All, I tried to join the second Domain for Ubuntu 22.04  and its not able to join the second domain by using this command "# realm join --user=ments.ldap domain2.com" and the output is --> realm: couldn't join realm: Already joined to another domain: domain1.com
[19:04] <Niclos98> I tried to join the second Domain for Ubuntu 22.04  and its not able to join the second domain by using this command "# realm join --user=ments.ldap domain2.com" and the output is --> realm: couldn't join realm: Already joined to another domain: domain1.com