=== WinterNight1 is now known as WinterNight [00:23] My computer’s HD was replaced, cloning the old drive to a new one. (It had a dual-boot setup - Ubuntu 19.04 and Windows 10.) I’m pretty sure GRUB was messed up, so I booted from the LiveCD to try and repair it but... [00:24] “sudo grub-install /dev/sda” < that’s the drive [00:24] gives me: [00:24] “grub-install: error: failed to get canonical path of ‘/cow’.” [00:24] What does that mean? [00:25] how do create a non lvm partition in the free space on /dev/md1? [00:25] parted/gdisk/fdisk don't see any partitions on /dev/md1 [00:25] https://i.imgur.com/9TgHebr.png [00:27] Got an error on removing apparmor here https://pastebin.com/XUay4v90 fail on snapd. [00:28] akem: please file a bug for that, ubuntu-bug snapd [00:29] Ok, i did the other one for evince. [00:29] removing apparmor .. not really a supported trick to do [00:30] yeah [00:30] but purging snapd ought not die like that [00:30] akem: "234 not upgraded." . What about ; sudo apt update ; sudo apt upgrade ; sudo apt -f insatll ' ? [00:32] this might be related https://bugs.launchpad.net/ubuntu/+source/evince/+bug/1798091 [00:32] Launchpad bug 1798091 in evince (Ubuntu Disco) "thumbnailer cannot create tempfiles (with apparmor denials)" [High,Fix released] [00:33] see comment #17 #18 and how to add yourself to the bugreport [00:33] Bashing-om, i fear something would break, i have a script to backup / in a tarball, i want to do that first. [00:33] evince-thumbnailer [00:34] I had some annoying issue with wifi drivers etc. [00:34] OerHeks, ok. [00:35] probably best to file a new bug and give that bug number in the description as a possibly related bug [00:35] well, not about that action to remove apparmor.. [00:35] not sure how one could fix that, besides a full backup [00:36] right, removing apparmor you're just kind of on your own. removing snapd ought to work, and that rm just looks busted, so a bug report there would be nice [00:43] OerHeks, but the bug report is for Nautilus, and the problem only occurs with Nemo. Anyway it's also related to some apparmor profile. [00:43] What are the risks removing apparmor? [00:45] no, it is against evince-thumbnailer [00:45] nautilius uses the same routine [00:45] Ah ok, yes. [00:46] SoItBegins: I can tell you that cow means copy-on-write: it's how a liveCD can be 'writable' without actually writing. So, /cow is something to do with the live CD. Run mount to check whether sda is the right drive; it may be that the names have swapped around [00:46] oke, this is disco, but comments say it is fixed, others say it did not.. [00:49] akem: apparmor confines services; we've got profiles on thumbnailers because they are easy to exploit [00:50] akem: we've confined a handful of services and applications with apprmor to try to mitigate the risks of those programs being exploited [00:50] akem: removing apparmor means those things are no longer confined. that might be fine if you never browse directories of pdfs or images downloaded off the internet, etc [00:51] akem: apparmor is also what confines programs in the snap packaging world; since that's an app-store style security model, it's what keeps the apps confined to just the plugs and interfaces that they need [00:53] sarnold, i see, thanks for the clarification. Yes of course i browse files downloaded from the internet and everything. Well i could still reinstall apparmor later and it should get back as before. [00:54] yeah [00:57] that rm -rf /var/cache/snapd/* in snapd.postrm shouldn't fail that way; what else has been done to that system? immutable? [01:03] TJ-, On my system? i didn't change much, no immutable anything afaik. [01:03] deltab: I figured it out. First, the EFI partition had only the windows boot information, so I added it back. That got GRUB working. Then, well… turns out the cloner didn’t respect partition order! So I booted into emergency mode and had to modify /etc/fstab with the new location of the EFI partition. [01:05] ah [01:05] BTW, there is one little issue left... [01:05] GRUB thinks my Windows partition is still at /dev/sda2, when the cloner moved it to /dev/sda4. How do I fix this? [01:05] edit the groub config? [01:06] Yeah, that. How do I do that? [01:06] I tried opening grub.cfg in the text editor, but there was a big warning at the top that said not to edit it manually. [01:06] SoItBegins: /boot/grub/grub.cfg is created by running "sudo update-grub" [01:06] Got it. [01:07] Did that. Seems to have worked, I think… I’ll reboot to double check. [01:07] SoItBegins: update-grub runs script from /etc/grub.d/* and reads /etc/default/grub for variable settings [01:07] SoItBegins: one of those scripts will call the tool "os-prober" that will discover the Windows OS and create the entry [01:08] It says the windows boot manager is on /dev/sda3 (the EFI partition) now. Testing it. [01:08] Windows 10 boots successfully as well. [01:08] SoItBegins: that is correct, on UEFI :) [01:09] Thanks for all your help! Now I can get to doing the stuff I wanted to do. [01:29] For those running on BTRFS, do you need to leave unallocated space for the snapshots to work or is that not needed? [01:29] Well, this is awkward. I’m trying to install the driver for a Wi-Fi dongle. [01:29] The thing is, the driver code is good through Kernel version 4.4. [01:29] I’m on Kernel version 5 and I’m getting a compile error. [01:42] gbit86: do you mean making snapshots, or reading from them? either way I'd make sure there's free space, because btrfs can't do much when it runs out of space [01:43] SoItBegins: what's the error message? === zbenjamin_ is now known as zbenjamin [01:44] deltab: It was that they changed the timer interface in kernel 4.15, so function “init_timer” was gone and an “__init_timer” was present. [01:44] I found a better driver for the dongle elsewhere. [01:44] Well I haven't ever really dealt with snapshots of any linux based fs before, so my only experience with such things is System Restore and VSS on Windows, which manages to create some sort of snapshot without requiring unallocated space. [01:45] I am guessing things are simply different in Linux, where knowing the actual details of what you have available up front is important. [01:45] ( old driver = https://www.tp-link.com/ca/support/download/tl-wn725n/ - new driver = https://github.com/lwfinger/rtl8188eu ) [01:46] And now maybe my connection won’t drop after the first 15 seconds. [01:46] From the research I have been doing I am partial towards just using lvm thin with ext4, after learning that I can't snapshot my regular LVM2 volume because I consumed the entire disk during setup -.-. [01:47] And sorry I didn't @deltab [01:47] there'd need to be some space to store the snapshot metadata, surely? [01:48] though it might not be very much [01:48] Right, so as long as I leave unallocated space it'll handle the creation of all the sub volumes I presume when I initiate a snapshot. === g is now known as glguy [01:49] I use VMs and docker a lot, so if I go the btrfs route I imagine I will need to disable the COW feature everyone keeps referencing as causing huge performance issues @deltab [01:50] a pity, because that's the kind of thing it's useful for [01:50] I guess it depends on how much RAM you have === [b0b] is now known as b0b [01:53] This guy says BTRFS does not require that I reserve space that the system is unaware of... https://superuser.com/questions/705304/reasoning-behind-hosting-virtual-disk-images-on-btrfs-filesystem [01:53] @deltab [01:53] 3rd comment down Thorsten Schöning [01:55] If that's the case.. and I can disable CoW on docker , VB and VMware directories then I will go ahead and try it, even with a performance penalty compared to other options. [02:13] gbit86: oh, you mean space outside of that allocated to btrfs? [02:15] If I install libasio-dev will that also pull in the rest of Boost? [02:17] leonardus: heh, it might; on my system, http://paste.ubuntu.com/p/5N2c7b3Fhd/ [02:17] :( [02:20] I updated the amdgpu-pro drivers on Ubuntu Budgie, and now I can't get past the login screen. The login screen is in a much lower resolution than normal, and when I log in, the screen goes black for a second before returning to login [02:22] I've had a similar issue before, and it seemed to be some sort of issue from the upgrade process, but I don't remember how I fixed it last time. I thought I just completely uninstalled the driver and reinstalled it, but that didn't fix it. And it uninstalled xorg in the process, so I had to reinstall that, though I think that's happened before, too === [b0b] is now known as b0b [03:55] Why is it that umask command has no effect inside a directory with chmod g+s set (setgid)? [03:56] umask 027 gives me file permissions of 640 on new files, but when I change into a directory with setgid, then new files I create/touch have permissions of 664. Why is that? [03:58] Can anyone hear me? [03:58] imsurit: Can you see my messages to the channel? [03:58] webdoll: yes, we can see your messages. [03:59] SwedeMike: Thanks! Just wanted to make sure. This is my first time using this web client. [03:59] Ubuntu 18.04 I believe apparmor is blocking rsyslog. In journalctl | grep denied I am getting "file '/var/log/syslog': open error: Permission denied [v8.32.0 try http://www.rsyslog.com/e/2433 ]" [04:00] in /etc/apparmor.d/usr.sbin.rsyslogd I added /var/log/** rw [04:00] -rwxrwxrwx 1 syslog adm 0 Jun 17 06:27 /var/log/syslog [04:01] can someone please tell me how I can enable apparmor to allow rsyslog writing to files in /var/log? [04:01] webdoll: i hear you [04:07] anyone? [04:12] hello, in ubuntu 18, what's the use of the second tray icon next after settings? https://pasteboard.co/IliziNZ.png [04:18] afidegnum: I think it is screen rotation [04:22] here i am.. on ubuntu reluctantly.. [04:25] potus: do you have an actual question? [04:30] I find myself running a docker script very often (https://stackoverflow.com/a/42419060) - but I'm always copy/pasting from this answer. Where should I install this little script so I can call something like docker-ls to call it? just install it in /usr/sbin ? [04:30] whois sam155 [04:39] I'm looking for the pypy twisted package for 18.04 LTS. [04:40] - haven't used Ubuntu w/ pypy yet, was kinda surprised not to find a pypy-twisted. [04:40] Hello. How to install exa? [04:41] I.e. https://github.com/ogham/exa ? [04:45] Also, how to install rust? Which package? [04:53] ryuo: do you? [04:54] potus: this is the ubuntu support channel, please focus on actual problems only [04:54] psprint: for git help, best to contact the git maintainer for help/support/install [04:56] ubuntu should switch to xfce for default windowmaker [04:56] potus: stop that please [04:57] also the ulgy off black color, why was that choosen? is there some significance? [04:57] xieta: apt-cache search twisted [04:57] lotussychje: sounds like your an efnet troll buddy.. rejoin #l*** with the rest of your mason friends.. stop trolling because once i make a move you'll get madd [04:58] now, i installed the os and im giving my comments to this channel which is suppose to have the developers [04:58] lotuspsychje, and then what? There are python-twisted-* packages but not pypy-twisted-* packages. [04:58] soo jump on another profile and say something dim witted. geez [05:01] xieta: we advice to install packages from the official repos as much as possible, if your needed package isnt there, we cant do much? [05:08] That's what I'm trying to understand - just curious why the official repos have pypy but not pypy-twisted. I'm also lurking in #pypy so when I learn what the recommended method for installation is - I'll share it here too. [05:08] Thanks for your support :) [05:21] I’m trying to install Vulkan on my computer. Does anyone have any recommendations on the most up-to-date way to do it? [05:30] SoItBegins, Have you tried Lutris first to see about it all working? [05:33] xieta: No. I don’t know what Lutris is. [06:16] I reinstalled apparmor, i'll try to just disable the evince profile for my thumbnails to shows up, it should be a bit better, less radical. [06:18] stuck here - I installed nvidia drivers for the 750m on this lenovo y510p via: ubuntu-drivers autoinstall - now I can't auto hide the dock for reasons I can't even explain, and even worse - apt is scaring me - https://paste.ubuntu.com/p/CKtgQKfF5M/ === morphis16 is now known as morphis === cupcake90 is now known as cupcake === imsurit_ofc is now known as imsurit [07:28] hi [07:29] anyone know how to remove bar border in conky ? been looking for that and did not find anything [07:38] brian1343: dont know about that one, but I use conky with lua rings like explained in https://www.unixmen.com/install-conky-lua-ubuntu-14-0413-1013-04-debian-fedora-linux-mint-opensuse/ [07:39] brian1343: you get quite nice stuff out from the conky with it. [07:40] pavlos: that's what i thought but my screen is not rotating. it's an hp Envy touch screen with screen rotation feature [07:41] ping goudberk.open.net [07:41] afidegnum: you have ubuntu on an Envy? How's it working ? Sorry missed the earlier conversation [07:41] sorry for that :) === imsurit_ofc is now known as imsurit [08:38] I’m trying to upgrade my Ubuntu graphics drivers to enable OpenGL 4.0 features. I have an AMD Radeon HD 6450 (Caicos). [08:49] Mava42: i asked a question earlier, " in ubuntu 18, what's the use of the second tray icon next after settings? https://pasteboard.co/IliziNZ.png" [08:50] afidegnum: press on it? see what it does? [08:58] hey, I'm looking at getting dropbear-intramfs setup to unlock full disk luks encryption remotely during boot, currently following this article: https://hamy.io/post/0009/how-to-install-luks-encrypted-ubuntu-18.04.x-server-and-enable-remote-unlocking/. I want to bond my ethernet and wifi so that there is a fallback if either fails, does anyone know if dropbear is able to work with this bonded interface? [09:00] afidegnum: how much ram memory you have? total. Just curious [09:37] j0h: it's not so much a matter of whether dropbox can do it, it's more a matter of whether the initramfs can. And I don't think it can (I could be wrong). [09:37] ahh gotcha tomreyn [09:37] I've searching for dropbear solutions but I should be looking for intramfs in my search I guess [09:37] you want to bond ethernet and wifi - how do you plan to do this? [09:38] well, I was hoping that it would be possible because it is possible on the full ubuntu os: https://lemmster.de/network-bonding-wlan-eth-on-linux-ubuntu.html [09:39] the idea is that, if the machine is plugged in remotely, it'll try and connect to either interface [09:39] just in case one isn't available [09:39] that was I can remote in and unlock luks [09:39] gofio: 16Gb [09:40] lotuspsychje: i did, i even rotated the screen, no success [09:41] gofio: no 12Gb [09:42] j0h: i see, i assume such a bond will require support on the other end of the connections. i just don't see this diuscussed. [09:43] hum, what do you mean by the other end? It would just be a simple active-backup which would switch if ethernet is not active? [09:45] j0h: you're right, it's not actual link aggregation, just a failover scenario, so i guess it doesn't need support on both ends [09:45] yea [09:45] j0h: i'm still not convinced that initrd will support it, though. [09:52] ahh ok [10:01] Can one also ask server/admin topics here or is this more for desktop users? [10:03] teasd: there's #ubuntu-server if you wanna be 100% sure [10:03] teasd: you could yeah, there's an #ubuntu-server channel as well if nobody here knows [10:03] That's probably a better fit, thank you! [10:10] Maybe gonna put it here as well, couldn't hurt: I have a server connected to our active directory via realmd/sssd. Problem is that everytime an AD-User connects, or when an AD-User runs the `groups` command, it says `groups: cannot find name for group ID 1234567`. This takes a long time everytime someone logs in via ssh. Does somebody know what the [10:10] problem could be? [10:10] yeah don't crosspost, though. [10:10] Didn't know, sorry :/ [10:20] Is the download server chooser a joke. It recommends for me to choose a server from USA even though I am in Finland. [10:28] On the other hand, http://mirrors.nic.funet.fi/ubuntu does not have e.g. usb-creator-gtk. What is wrong with it? [10:29] afidegnum: thanks. Not bad! [10:30] enial-security/main amd64 to be more specific [10:43] jarnos: it is recommended to get security updates directly from security.ubuntu.com for exactly this reason - so that you get security updates immediately [10:52] thsnr, then software-properties-gtk should setup security updates to download from there and not change them to be downloaded from other mirrors. [11:14] Howdy folks [11:32] Morning [12:31] hi, how can you know all the servers running on the system ? [12:34] Cooler, "know all the servers" ? Do you mean services or something else? [12:36] Cooler: virtual machines? [12:36] Cooler, if you do mean services, you might be able to use service --status-all [12:36] services like apache dns [12:36] ect.. [12:36] i mean :) [12:36] Cooler: ha :) [12:37] Cooler, if using systemd.... systemctl list-unit-files | grep enabled and a little netstat -tunelp wouldn't hurt either [12:38] ok lets try that ! [12:39] Cooler: you need net-tools on latest systems or use ss [12:41] i have 2 monitors connected via DP that are capable of 1440p and 144hz... however it i'm unable to set them to 1080p. 1920x1080 is not recognized by the system. i'm trying to add the resolution in xrandr with no success, any help? [12:42] https://paste.ubuntu.com/p/GPBs4HptcY/ [12:52] inascime: https://superuser.com/questions/119755/hdmi-with-resolution-2560-x-1440-possible perhaps? [12:54] i'm able to use 1440p . the problem is i want to use 1080p resolution but it isn't recognized by the system === waveform_ is now known as waveform [13:00] I've got a problem with a Mellanox Bonded port dropping out, I've changed the DAC cables, upgraded the firmware on the cards. I'm about to install a 5.1 kernal on 18.04 to see if the kernal driver is better. I'm really at a loss. I have the same card versions on other servers with the network set the same way and I've not had any issues. But I have 4 servers that randomly start dropping their ethernet links. Ideas Ping me please. [13:02] If I reboot the server it takes a few days before it will start flapping again. It's causing problems in my ceph cluster [13:05] mTeK, what firmware do you have installed? [13:06] firmware-version: 2.42.5000 [13:06] It was on 2.40.XX and I flashed the new firmware to try to stop this. [13:08] The cards are connectx 3 [13:08] Does dmesg give any useful info? I can't imagine it'll be a hardware issue if it's happening across various hosts [13:10] Not that I can see. https://pastebin.com/mfMqMVcJ [13:11] I thought it was the DAC's so I changed those out. I originally had quanta switches and then moved to arisia, I know the DAC's are different so I changed them [13:11] These servers are mlaged btw [13:12] I have 7 other servers, 3 with 18.04 with the same network config and 4 with debian and ovs not having the issue. [13:14] Same Nics connectx3 might have a pro in there someware. The only difference is the 4 that are having trouble are supermicro servers and the others are dells. [13:19] I'm wrong these 4 have the 02:00.0 Ethernet controller: Mellanox Technologies MT27520 Family [ConnectX-3 Pro] [13:27] cupcake ? [13:27] are you there ? [13:27] I have a script that when executed as "php myscript.php", will run much faster than if I run it in the background with "nohup php myscript.php >/dev/null 2>&1 &". All this script does is execute a task, sleep for 10 seconds, then run again and it should be running all day long. However, the difference between the 2 methods is probably about 35 seco [13:27] nds between executions. For some reason the background process sleeps far longer than is defined in the script. [13:27] "net-tools on latest systems " i installed that net-tools but i dont know how to use it [13:28] im googlingit :) [13:30] Cooler: like cupcake90 said, netstat -tunelp. this will show you all programs that are listening on TCP or UDP sockets [13:31] sudo netstat -tunelp, to see the names of all programs [13:36] just upgraded to 19.04 and I'm having an issue in intel gfx where my display hangs or blacks out briefly when fullscreening a video from a browser or a video player [13:41] hey, is this an appropriate channel for ubuntu core distrib help? [13:44] ooohhh ok === sparty_ is now known as sparty [13:50] amdgpu-pro-18.20-606296 doesn't work with 4.15.0-52-generic, so i tried updating to amdgpu-pro-19.20-812932 , but with that the system doesn't recognize my RX580 at all [13:51] sesors can't see it [13:51] but if i uninstall the driver then sensors can see it [13:53] .. i will try to install 18.20 again just to double-check that it's not possible [13:56] I had this issue with my Radeon Vega Frontier Edition. Ended up just using the built in default. [13:56] It wouldn't sense anything plus it made the video even more choppy. [13:57] Error! Bad return status for module build on kernel: 4.15.0-52-generic (x86_64) [14:07] port 53 and port 631 are on listening mode. is it normal ? [14:09] Cooler: port 631 is for CUPS, which is what allows other devices to send print jobs to the printer. Normally both of these ports are bound to the loopback interface so the local machine can use them [14:09] oh ok thanks pragmaticenigma [14:11] Cooler: As long as those ports aren't exposed or accessible to the Internet you will be fine. If you have configured your modem/router to forward those ports or set that computer to the DMZ, you are at risk and should disable those options. [14:12] i understand that now :) pragmaticenigma [14:14] ty [14:14] the DMZ is for testing isnt ? [14:16] or redirecting all the packets to one machine [14:17] DMZ is for avoiding the firewall, Cooler [14:17] ohh [14:19] ok i understand now [14:19] bad name, demilitarized zone... when it's basically the opposite, it's the war zone [14:20] if an intruder get his hand on your router and set it to DMZ some local ip he can attack any machine ? [14:20] " the war zone " [14:20] Cooler, if intruder gets their hand(s) on your router, you have already lost the attack [14:21] Cooler: No, DMZ means that any traffic sent to your router is passed directly to the computer defined in the DMZ. Meaning that if that computer does not have a firewall of its own, it is completely vulnerable to attack from the Internet [14:21] But, yes, basically... if they have control over your router, they can attack any system the router talks to... but, that system still has to have vulnerable services /etc [14:23] I see... [14:23] would you use a machine with iptable and then configure other systems to connect trought that machine with DMZ ? [14:26] that s why you have to patch your router with the latest firmware ? [14:27] im noob :) [14:28] JimBuntu, if you loose the attack. What can you do shutdown the router ? [14:28] Cooler: pull the plug? [14:29] only real option if you've lost control. [14:29] I'm using openbox and lightdm on xenial. openbox says it has no power management by default. xscreensaver is set to mode: off. what else could have just switched off my screen? [14:30] Sven_vB: X has a timeout where it'll blank the screen after a period of inactivity. see xset. [14:30] ryuo, thanks! [14:30] dpms [14:31] that s right ryuo [14:31] xset q says Screen Saver: prefer blanking: yes allow exposures: yes timeout: 600 cycle: 0 [14:31] DPMS (Energy Star): Standby: 600 Suspend: 0 Off: 900 DPMS is Disabled [14:31] so it's probably the screen saver timeout? [14:32] Sven_vB: both i guess. i never looked at it that closely. [14:32] Sven_vB: just know it has ways to configure it when not using a full DE. [14:33] ok [14:33] afaik this is primary a feature of X and not the DE. [14:33] primarily [14:33] DE just configures and hooks into it. [14:36] I've just added sh -c 'xset -dpms; xset s off' to my autostart and will see if that helps. [14:37] also ran it right away and now my screensaver timeout shows 0 so probably that was it. [14:48] Hi! I have dualboot Win7/Kubuntu laptop with partitions /dev/sda1 – 5 + an unallocated/unknown partition. /dev/sda4 it the extended partition that holds /dev/sda5 and /dev/sda6 + the unallocated one. (/dev/sda1-3 are windows 7). Now I need to delete the /dev/sda5 (old ubuntu studio install) and use the space from it + the unallocated one in /dev/sda6, the active Kubunu. It’s not possible for med to delete /dev/sda5. Can someone help [14:48] me with that? ..Grub must also be cleaned up.. [14:49] I Asked in #Kubuntu and was told to boot live USB - do what I want to do - and then chroot and update grub. Was not quite comfortable with that reply since I am new to linux... [14:51] Thr0r: Please don't cross post, and if you're not comfortable with the procedure, then ask the helper in the previous channel to step you through the process. [14:53] OK - Thank you so much [14:55] should I use this netplan to setup my networking continue to use interfaces? [14:55] ubuntu server 18.04 [14:58] j0h: come join to #ubuntu-server please [14:58] thanks lotuspsychje [15:02] if I install zsh, will it automatically move my all of the stuff in my bashrc to zshrc? If not, how do I sync my settings then? [15:05] pragmaticenigma: Could you at least confirm somehow that the advice I got was somewhere near what I should do? [15:05] I find myself running a docker script very often (https://stackoverflow.com/a/42419060) - but I'm always copy/pasting from this answer. Where should I install this little script so I can call something like docker-ls to call it? just install it in /usr/sbin ? [15:08] jayjo: ~/.local/bin (and make sure that's in PATH) or /usr/local/bin (if you need it system-wide). you could also set shell aliases for it. [15:09] should I use `install` or just move it manual? [15:10] Thr0r: If you're not comfortable with the advice given, then don't do it. Simple as that. If you're that concerned then you should have Disk image backups to make sure you can recover from any mistakes. [15:11] Thr0r: The volunteers here are very knowledgeable, but can make mistakes. It is ultimately up to you to decide who and what you trust, and for you to do the research necessary to verify what you are being told. [15:13] having backups helps when things go wrong [15:14] pragmaticenigma: OK - I was sure this was the support channel where one could trust advice and not get into problems, I am new and is not in a position where I can trust everyone just like that. Backups I have but time I don't.. [15:16] Thr0r: If you don't have the time to devote to these projects, then you should reconsider what you are doing. [15:19] pragmaticenigma: I just wanted to give Linux a chance.. but seems I will go back to Windows pretty soon.. [15:19] !dualboot | Thr0r start here [15:19] Thr0r start here: Dual boot instructions: x86/AMD64: https://help.ubuntu.com/community/DualBoot/Windows - Macs: https://help.ubuntu.com/community/MacBookPro https://help.ubuntu.com/community/YabootConfigurationForMacintoshPowerPCsDualBoot === rikmills is now known as RikMills [15:22] Thr0r: We're all here to help, but it is up to you to decide what you're comfortable with. I know I have talked to you before and suggested that to try out linux, consider using a virtual machine instead to test your ideas out in. We are all volunteers here, and we all have limited time to be able to help. To request that someone do all the validation and verification that you are capable of doing on your own isn't fair to the [15:22] rest of members of any channel seeking help. It monopolizes the time and attention that others also need. [15:24] pragmaticenigma: I just wanted help to remove a partition - thats all. But thanks for all help. [15:40] i'm trying to share my wifi connnection to a device via ethernet, i went to the connection editor and set the ethernet connection's ipv4 settings to "shared to other computers" and it seems to be working, but i can't figure out what ip address the device has via connection sharing, anyone have an idea how to do that (or test if it's working)? [15:41] mmazing, you could run "ip addr" and look for an IP on the ethernet [15:42] i see one, but when i go to it it's just my computer and not the device [15:42] inet 10.42.0.1/24 brd 10.42.0.255 scope global enp4s0 [15:42] i nmapped 10.42.0.0/24 and can't find it that way either [15:43] although when i nmap that range the ethernet activity lights up on the device [15:47] mmazing, try netstat -tap [15:49] BluesKaj, only thing i see there related to that network is 10.42.0.1:domain - dnsmasq [15:49] seems like it's not getting an ip address [15:51] does this exist somewhere - ported/forked version of 18.04 for powerpc (g4,g5) - not looking for quick answer just would like to know if a ubuntu ports exist like it does for debian? [15:52] not 18.04 for higher than g5....it's nice to see power8 but I don't have that hardware [15:52] going back to lurk mode [15:53] !mac | donofrio [15:53] donofrio: For help on installing and using Ubuntu on a Mac, see: https://wiki.ubuntu.com/MactelSupportTeam/CommunityHelpPages [15:53] donofrio, and... https://help.ubuntu.com/lts/installation-guide/powerpc/index.html [15:56] again, thank you for the quick links but I'm looking for current revision 18.04 or 19.04 in a "Ports" version like I have with Debian 10 (granted while not supported it does work) so I'll go back to lurk and I thank you for considering this request/question [15:57] for PowerPC G4 and G5's do not have anything newer than that in PowerPc Arch [15:58] mmazing, try arp -a [16:01] BluesKaj, no dice, just other stuff on wifi network, should i be using a crossover cable? [16:01] hey do you remember what was your favorite retro gaming game ? [16:01] donofrio, is this not the installer you need? http://ports.ubuntu.com/ubuntu-ports/dists/bionic/main/installer-ppc64el/current/images/ [16:01] sorry wrong channel [16:02] mmazing: the NetworkManager dnsmasq/dhcp server running on the shared connection will issues leases in the 10.0.42.0/24 range [16:03] mmazing: you should be able to find the leases files under /run/NetworkManager/ or possibly /var/lib/NetworkManager/ [16:03] donofrio: If it isn't listed on the ISO download screen, it's not supported, and therefore no one here would have any knowledge or links to send you to find it [16:10] donofrio: it was finally dropped in 2016, see https://lists.ubuntu.com/archives/ubuntu-devel-announce/2016-December/001199.html [16:11] mmazing, I thought you were using a cable [16:13] say, for instance, you had a server 400 miles away and no console access, and screwed up the ~/.ssh/config file for a user, is there any way to remotely ignore that syntax error and log in anyawy? i almost certainly know the answer is no, but tt's worth checking before spending money [16:15] Cheez, I don't have an answer for that, however if no one answers here you could try the #linux channel as well [16:16] what's the best way to sync up a google drive with ubuntu 1804 LTS. I am trying to sync my KeePassXC database across a few hosts - seems to be the quickest way [16:16] BluesKaj, i am using a cable, but it's a normal ethernet cable not a crossover cable [16:18] coz_: i'm pretty much resigned to paying the remote hands fee at this point. :) [16:18] Cheez: The answer is no; unless your service provider offers another remote management option, or you have effectively locked that user out of the system. If another user account has access, and has sudo privileges, that user could restore the damaged file. [16:18] i'm almost certain there's no way to bypass it [16:18] odd thatir doesn't show up, unless it's part of your ethernet name and circuit [16:18] no other user has ssh rights :) [16:18] mmazing,^ [16:18] Cheez: then you are out-of-luck i'm affriad [16:19] my own stupid fault really, i did some work on it, i saved and left the session open, and my connection dropped :D [16:19] didn't create a backup user just in case [16:21] the server has an ILOM card, but the card uses https only, is signed with an MD5 cert and requires java 6, which essentially means i'll have tp create a windows XP VM and use IE to even load the damn thing [16:21] no current browser will let you even touch it [16:22] (in normal use the network port the ilom card is plugged into is disabled, ironically i can enable that remotely fine!) [16:27] mmazing: gig-e ports will do the crossover - no need for special cable any more. [16:27] Cheez: unless you run other services which intentionally or unintentionally allow for remote administration of this system there's nothing (except for the ILOM) and a (just) potentially configured and physically existing serial console you can use / do. [16:27] !server | Cheez [16:27] Cheez: Ubuntu Server Edition is a release of Ubuntu designed especially for server environments, including a server specific !kernel and no !GUI. The install CD contains many server applications. Current !LTS version is !Bionic (Bionic Beaver 18.04) - More info: http://www.ubuntu.com/products/whatisubuntu/serveredition - Guide: https://help.ubuntu.com/lts/serverguide - Support in #ubuntu-server [16:28] Cheez: actually, ~/.ssh/config is the ssh client configuration, it is not involved in ssh server operation. [16:29] i.e. whether this file is broken, exists or not, should not matter for whether you can authenticate to this systems' running ssh server. [16:29] top/htop shows some processes using about 80% of all 4 cores. (good) -- I need to email this to someone, but screen shots and cut/paste console output is icky. is there something that will format that info in a cut/paste/email friendly format ? [16:30] CarlFK: That isn't entirely true. The ability of the port to do cross over depends on the board manufacturer, not the network speed capabilities. Most newer cards auto negotiate the cross over, the only way to know for sure is to consult the network cards documentation === subopt is now known as _subopt_in_repos [16:30] Cheez: this file, on the remote server, that is. if you broke it on your local computer and thus lost information on how to connect to and authenticate against the remote server, this could be an issue. [16:30] pragmaticenigma: it is in the spec. if it doesn't, then technically it isn't gig-e. [16:31] which is kinda circular, but meh. Ill bet the number of cards that dont' do it are small enough to ignore. [16:32] CarlFK: "man top" tells you that argument -b will generate output that can be redirected to a file or other destination. Used with "-n 1" will report one cycle of data [16:32] doh. my answer was right under my noise. [16:35] CarlFK: If you can provide a link to the documentation that state Gig-E always supports "auto-mdix" I would be interested in reviewing it. All the documentation I have found says it can be enabled, but nothing about it being required. [16:35] Cheez: you mentioned the ILOM card uses old cert MD% etc; in that case I use an old 32-bit build of Netscape Navigator [16:40] pragmaticenigma: i'm parroting bwh, a DD who's full time job was linux nic drivers for a company that sold 10g nics. [16:41] okay, i'll keep searching for something to add to my library. thanks for the heads up CarlFK [16:47] pragmaticenigma: on my top thing - how do I show multiple cpus (what I get hitting 1) in the -b mode? [16:47] 1 :Single/Separate-Cpu-States toggle [16:51] pragmaticenigma: IEEE 802.3ab (1000BASE-T) makes auto-MDI/MDI-C an optional feature not a requirement [16:51] grrr MDI-X ... silly fingers [16:53] That's what I thought TJ- ... so CarlFK's information isn't fully correct. Gig-E can optionally support auto crossover, its not a reguirement? [16:53] pragmaticenigma: it's more complicated than that :) [16:54] okay, I'll leave it at that then. Thank you both for the information [16:54] pragmaticenigma: for 1000BASE-T in most case auto-MDI-X is not required [16:55] pragmaticenigma: usually the phyical layer (PMA) detects which way cables are wired even when the pairs are wired "wrong" [17:00] Hi all, i am trying to get openvpn server going on ubuntu 18.04 lts; what has happened to easy-rsa in the packages? [17:02] roy2098: what do you mean by "what has happened to easy-rsa in the packages"? [17:02] !info easy-rsa bionic [17:02] easy-rsa (source: easy-rsa): Simple shell based CA utility. In component universe, is extra. Version 2.2.2-2 (bionic), package size 16 kB, installed size 93 kB [17:03] tomreyn: Package easy-rsa is not available, but is referred to by another package. [17:03] !universe | roy2098 [17:03] roy2098: The packages in Ubuntu are divided into several sections. More information at https://help.ubuntu.com/community/Repositories - See https://wiki.ubuntu.com/RecommendedSources for the recommended way to set up your repositories. [17:03] ok will do thanks! [17:05] my, that's an old how-to i pointed you to [17:05] https://help.ubuntu.com/community/Repositories and https://help.ubuntu.com/community/Repositories/Ubuntu are better [17:06] ...than the "RecommendedSources" page [17:06] tomreyn: here's what is in my sources.list - deb http://archive.ubuntu.com/ubuntu bionic main [17:06] i've updated and upgraded [17:06] still no easy-rsa [17:07] roy2098: as indicated by ubottu, easy-rsa is in "universe" [17:07] you only have "main" so far [17:09] ah, thanks again! [17:10] roy2098: note the different support policies for these repositories as outlined on the wiki i pointed you to [17:10] CarlFK: i can't find anything in top to get the cpu utilization [17:10] CarlFK: I thought maybe nmon might have something, but that also is a dead end [17:11] htop doesn't appear to have a mode for an external output [17:12] pragmaticenigma: k - thanks for trying - "we" (or they?) are trying to figure out how much cpu is needed to run a set of processes - "we need i7!!!" um.. I use i5s and I dug up an i3 that seems to be working (but one process shows 97%.. [17:13] ps auxf > file.txt [17:13] i3-2120 from 2009 - anyone know what gen that is? [17:14] probably the first [17:14] CarlFK: gen 2... the first digit indicates the generation [17:14] ioria: I need to show the total [17:15] sandy bridge [17:15] CarlFK, awk and sum it up ? [17:15] sandy bridge sounds familiar. [17:15] man top [17:15] ioria: too hard. TJ: that doesn't prove anything ;) [17:16] ok [17:22] CarlFK: huh? [17:23] TJ-: They are looking for a capturable output of CPU utilization. top doesn't appear to provide that in an exportable [17:23] way [17:24] top -b -n1 > /tmp/top.log [17:24] I have the box here and can see the results: looks OK. trying to document that [17:24] tomreyn: it doesn't break out the cpus like top ... press 1. [17:25] pragmaticenigma: "sar" [17:26] !info sysstat [17:26] sysstat (source: sysstat): system performance tools for Linux. In component main, is optional. Version 11.6.1-1 (bionic), package size 297 kB, installed size 1224 kB [17:28] proper monitoring would work, too ;) [17:29] TJ- like you, I am only willing to put so much effort into helping someone else ;) [17:31] when my script mounts an ext3 partition read-only into a mount point that is used only by my script (e.g. in its temp directory) and later umounts it, will this interfere in some way with other users mounting the same partition, possibly even rw, as long as my script and the interactive users don't read/write the same files? [17:31] the top (snicker) problem is: DebConf video team needs 3+ boxes in about 4 weeks. they think they can afford to hold out for i7s. https://wiki.debian.org/DebConf/19/VideoTeam/VideoEquipHunt [17:32] CarlFK: and the debconf video team runs ubuntu? [17:33] tomreyn: they did the year I was the lead ;) (and yes, there was a bit of grumbling) [17:33] :) [17:36] CarlFK: you won't normally be able to rw-mount this file system multiple times. so a user assuming they could do this would be interfered with. [17:37] tomreyn: um... wrong nick I hope [17:37] CarlFK: sorry, not you, this was for Sven_vB [17:38] Sven_vB: oh you said your script mounts read-only. i guess it can be ok then [17:39] the other user may still run into a warning or error even, i'd recommend testing it [17:42] can you PLEASE make the ubuntu updater stop popping up in my face? i will do it by terminal when and if i choose [17:43] sure, let me quickly login to your system and reconfigure it for you. [17:43] done! [17:43] well, jokes on me i suppose [17:43] but yeah. maybe you could share with me how to configure that? [17:43] Gallomimia: you can just remove the update manager and such [17:43] remove it? [17:43] oh my i didn't know that [17:44] you could always go to Mint where the updater doesn't nag you. :D [17:44] lol [17:44] tomreyn, thanks! [17:44] sudo synaptic [17:44] apt-get remove update-manager ?? [17:45] whoa. it's going to remove ubuntu-desktop ? this doesn't seem wise [17:45] and perhaps remove gnom at the first apt autoremove [17:45] Gallomimia: update-notifier-common [17:46] Gallomimia: and ubuntu-desktop is just a meta package to help you get (and keep, if you wanted) the standard ubuntu desktop installed. [17:48] hey how do you share your ubuntu drive to a win 10 machine ? [17:48] Cooler: you don't. Win10 can't speak ext [17:48] dual boot or over the network? [17:48] over the network [17:48] local [17:48] samba [17:48] Gallomimia: you could also choose to have all updates installed automatically without you ever getting prompted. that way you get no notifications and still benefit from timely security updates [17:48] I thought they are some ext reader in the win world [17:49] carefully. [17:49] samba ok ! [17:49] thanks ! [17:49] tomreyn, that is an ass-biter if i ever seen one [17:49] careful with the samba install [17:49] the last time i did it, i picked the wrong version and everything got borked. i couldn't remove it safely either [17:50] is there an easy way to schedule a mount point to be umounted as soon as no process uses it anymore? [17:50] Sven_vB: automount ? [17:50] well... i will use a usb3 external drive then [17:51] if you have one of those and you're just copying a few files over, that's probably easier [17:51] you can also use putty and similar to copy files via ssh and scp [17:51] Cooler: not sure if it's work: https://www.howtogeek.com/112888/3-ways-to-access-your-linux-partitions-from-windows/ [17:52] Ool, not sure what you mean [17:52] that's for dual boot [17:52] I mean I don't use windows :) [17:52] ool he is dealing with 2 separate machines [17:52] Ool, in my case the mount point is in the temp dir of my script and ideally would also be rmdir-ed after it's umounted. I'm trying to avoid litter in case my script crashes. [17:52] wants to share over network [17:52] 1 ubuntu and the 2 with win 10 [17:53] sorry… winscp to see it across ssh [17:54] that's a much more useful and secure way to do it than samba [17:54] there's also SSHfs for windows https://github.com/billziss-gh/sshfs-win [17:54] altho, slightly less convenient than having it mounted permanently. depends how long you want shares to be available [17:55] can i just go ahead and say i'm really disappointed with the "progress" that's been made in local area file sharing in the past 28 years? [17:57] not here, since that's a discussion topic [17:58] Gallomimia, let's discuss in #ubuntu-offtopic what features you see lacking in sshfs ;) [18:04] ok [18:08] next question: anyone know a decent script for purging duplicates from a music collection? turns out all my music files are literally in there twice. i'm using rhythmbox to browse and play, hearing every song twice is pretty lame [18:09] Gallomimia, if the files are bit-for-bit identical, not just the same song, you can use 'jdupes' [18:10] where are the dupes located relative to one another? [18:10] that's a big "who knows" buddy [18:10] there's a few nested folders in my /data/Music/ dir [18:10] Gallomimia, I'd just run jdupes on the whole drive. [18:11] If you find a bunch of matches in the same place, ^C and delete the whole folder, then run again. It doesn't match whole dirs [18:11] as an added note, a few of them are actually missing chunks out of them. shorter than their dupe and skips mid-song [18:13] reading about on github ty [18:13] Are snap packages architecture agnostic? As in, if a snap is listed on snapcraft, does that mean I can install it the same on my x86 or ARM systems? [18:16] lordcirth: fail: https://pastebin.ubuntu.com/p/vYby3DcwSp/ [18:16] Gerowen: I think that's up to the snap authors, they'd have to build multiple versions [18:16] leftyfb, that's weird [18:17] I get it when running on my home directory as well [18:17] leftyfb, 18.04? can you strace it? [18:17] Gerowen: not necessarily agnostic, there are build configurations. /join #snapcraft / #snappy [18:18] lordcirth: https://pastebin.ubuntu.com/p/8tQyR5XzFb/ [18:19] leftyfb, what filesystem? Does it always segfault on that file? [18:19] ext4 [18:20] let me check [18:20] it does. It's a directory with 22 video files in it [18:21] all under 1G in size [18:21] leftyfb, can you try the older 'fdupes'? [18:21] jdupes is a fork that is way faster [18:24] leftyfb, if you remove the 'A' option ? [18:24] fdupes runs without error. Though I'm not sure how it's supposed to work. It listed out 633(out of 1021) files in the same directory. They're all timelapse pictures so maybe it considered half of them the same picture? [18:25] ioria: tada! [18:25] leftyfb, if the hash matches, it then does a bit-for-bit compare. [18:26] lordcirth: the bits shouldn't be exactly the same really. At the very least the timestamps are different [18:26] ioria: ah, I didn't mean to use -A anyway. I misread the man page and meant to use -1 [18:26] leftyfb, ok [18:26] though it really shouldn't segfault with -A [18:27] of course [18:28] leftyfb, jdupes 1.9 on 18.04? [18:30] correct [18:31] leftyfb, that's weird. Could you please file an issue? https://github.com/jbruchon/jdupes/issues [18:32] can't at the moment. I was just trying it out of curiosity. I'm in the middle of some things at work and might have to file some bugs with Intel instead :) [18:34] Ok, no problem [18:37] hello [18:37] FreeBDSM, hi [18:39] a general linux question: how to echo an arbitrary process's envs? [18:40] FreeBDSM: cat /proc/pid/environ [18:41] sarnold: hmm, thanks, keep forgetting about /proc/ [18:42] FreeBDSM: there's loads of neat things in there :) [18:42] sarnold: like what? [18:43] how to cause a hard system reboot by echoing something to /proc/? [18:43] just came home from a job interview where I got asked this question [18:44] FreeBDSM: /proc/pid/exe is the executable that's running a process.. once, when I upgraded screen, I couldn't re-attach to an existing session because the protocol had changed between versions. but it was easy to use /proc/pid/exe to execute another copy of the OLD screen binary, to then reattach :) [18:44] That's a fun question [18:44] /proc/sysrq-trigger [18:44] fresh install of 19.04, installing updates, it hangs on "Configuring snapd..." with the last console message showing "snapd.snap-repair.service is a disabled or a static unit, not starting it." - any idea? [18:45] sarnold: thanks [18:45] `echo 1 > /proc/sys/kernel/sysrq && echo b > /proc/sysrq-trigger` [18:45] only as root. Otherwise, use tee [18:45] sudo tee [18:46] yeah, I learned how to sudo tee [18:55] sarnold: can I safely modify /proc/${pid}/environ to add a new env? [18:55] FreeBDSM: no, it's read-only. you could use a debugger though [19:12] sarnold: could you explain a bit more in detail about re-attaching to the old 'screen' process. What happens when you run old screen, then update screen package so it replaces the file, but the process of old screen still runs. Does system copy the old executable right into /proc/${pid}/exe (because usually that's just a symlink to the regular executable file of the program)? [19:12] or does the system move the old executable just elsewhere (and where? /tmp/?)? [19:14] also, I'm exploring /proc/ now and I see that there's /proc/${pid}/root which is a symlink to '/', is this how chrooting works? === Alina-malina_ is now known as alina-malina [19:14] FreeBDSM: on unix-like systems it's possible to unlink any file, while it's in use; the file data will stay on disk until all processes that have it open, close the file [19:15] sarnold: but how does that work? where will /proc/${pid}/exe point? [19:15] I lack some knowledge about interconnection between files on filesystem and inodes, I guess what you meant that unlinking a file doesn't unlink it's inode [19:15] here's an example from one of my shells: [19:15] lrwxrwxrwx 1 sarnold sarnold 0 Jun 26 18:35 exe -> '/bin/bash (deleted)' [19:16] Hi! What does 'nosplit' exactly mean in apt? [19:16] sarnold: so how did you use it then? [19:16] FreeBDSM: just a ./exe :) [19:16] FreeBDSM: a few lines output: http://paste.ubuntu.com/p/DzVVnTXwMV/ [19:17] sarnold: do I get it right that you could basically restore the executable via `cat /proc/${pid}/exe > myexecutable.exe`? [19:17] Xalys, i'd chck my /etc/apt7sources.list [19:17] ioria: that's not the issue [19:17] I changed web servers on the apt rep [19:17] o [19:17] FreeBDSM: yes. exactly! :) http://paste.ubuntu.com/p/tGk2JXNQjS/ [19:17] But there doesn't seem to be any docs on what the error means exactly [19:18] Xalys, can you paste 'sudo apt update' ? [19:18] sarnold: but it's quite rarely when a program is packed just as a single executable, it usually relies on libs and stuff, do they get deleted? [19:19] ioria: E: Failed to fetch http://xxx/dists/stable/InRelease Clearsigned file isn't valid, got 'NOSPLIT' (does the network require authentication?) [19:19] Xalys, you know it's not an ubuntu source ? [19:19] ioria: yes, it's my apt repo [19:20] FreeBDSM: yeah, if those get deleted, this trick might not work so well [19:22] sarnold: thank you a lot for explaining such things, I appreciate it a lot [19:22] FreeBDSM cool :) [19:22] I like explaining things, too, hehe [19:23] I'm exploring /proc/ now and I see that there's /proc/${pid}/root which is a symlink to '/', what if change the link to elsewhere? is this how chrooting works? [19:23] sometimes people don't always like explanations though, lol [19:23] that's another read-only thing [19:23] it'll just indicate where in the filesystem the process's root is [19:24] sarnold: how come it's read-only if it's mode is lrwxrwxrwx? [19:25] FreeBDSM: all symlinks on linux are always 777 [19:25] (well, if they come from another unix system, they might not be. it's weird.)( [19:26] ah, I see [19:26] and chmoding the link chmods the file it points to [19:26] weird design, IMO [19:26] why not mirror target's mode? [19:27] because then people would need to manage it :) [19:27] manage how? [19:27] also, the file it points to can be swapped out, mode changed, etc, and updating any symlink permissions at the same time would be very expensive [19:27] right [19:27] that's a good explanation [19:28] "I can't read your symlink can you please change the modes on it so I can see what the symlink points to" etc [19:28] alright, time to prep for a meeting.. have fun FreeBDSM :) [19:28] if you can't read the symlink - means it's target would have such a mod that you don't see that file :) [19:29] sarnold: okay, good luck with the meeting and thanks for help [19:32] ioria: issue found. apt does not like it when the content-type header is missing. I have to say the docs on this really suck. [19:32] Xalys, ok, good [19:33] JimBuntu, no that one threw a open firmware error http://paste.opensuse.org/39564924 [19:55] Hey. I installed the golang-1.12 package but cannot run go commands. I've checked again and it shows it has been installed so I don't get what could have gone wrong. Any help? [19:57] sometimes you need to log out and log in for the path to work right [19:57] I would try doing that first. [19:58] i forgot to add. I have rebooted as well just to be sure. And still no luck [19:59] Did you install this package from the ubuntu repos? [19:59] via apt? [19:59] yup. specifically the golang-1.12 package [19:59] what command are you trying to do? [20:00] I usually check using "go env" [20:00] and i get the message "command 'go' not found" [20:01] can you do "man go". doo you get a manual page? [20:02] hil8: sudo apt install command-not-found; go [20:02] zapproger: just tried it and there's no manual entry for it [20:03] well, disco has go 1.10.4 [20:04] hil8: which ubuntu version are you running there? [20:04] tomreyn: i'm a bit confused. should it be sudo apt install command-not-found? I'm on 19.04 [20:05] hggdh: yup that's the default here as well. But I was hoping to get something newer than 1.10 [20:05] hil8: yes, i you'll sudo apt install command-not-found and you'll run "go" again it should tell you which package you need to install [20:05] !latest | hil8 [20:05] hil8: Packages in Ubuntu may not be the latest. Ubuntu aims for stability, so "latest" may not be a good idea. Post-release updates are only considered if they are fixes for security vulnerabilities, high impact bug fixes, or unintrusive bug fixes with substantial benefit. See also !backports, !sru, and !ppa. === nshireTimeout is now known as nshire [20:07] tomreyn: it lists golang-go and gccgo-go which are both 1.10. There's also a snap package which is 1.12 [20:07] If all fails i might try that [20:08] a !ppa or !snap is probably the right thing to do (if you can find one which looks like you should be able to rely on it) [20:09] ...assuming you really need the newer go version [20:10] there is a snap with go 1.12, from mwhudson [20:10] trying out webassembly and 1.11 or newer is required. I'll give one last go(reinstall) and see how it goes [20:10] hggdh: yup. i might have to go with that [20:25] ll mloza [20:26] Ubuntu encrypted install, stuck at initramfs. Never seen this. Booting from a USB I can mount the drive with the password and everything is there, just can't boot normally. Screenshot https://i.redd.it/wsvi4nobk7631.jpg I have a thread on the forums here https://ubuntuforums.org/showthread.php?t=2421770 if you would prefer replying there. [20:27] intok: Not having seen your screenshots, but I’d guess `cryptsetup` is missing in your initramfs. [20:28] intok: Sometimes that’s ignored while building the image (because misconfigured /etc/crypttab). [20:29] intok: `CRYPTSETUP=Y mkinitramfs ...` from a USB booted chroot environment should fix this temporarily. [20:30] vlt try it via reddit post I made https://www.reddit.com/r/Ubuntu/comments/c4gf1x/getting_initramfs_on_luks_encrypted_system/ [20:31] intok: Try what? [20:32] vlt mdadm keeps looking for arrays? The screenshot, thats where I am trying to link it from [20:33] intok: What does `cat /proc/mdstat` say? [20:33] intok: And `/proc/cmdline`? [20:34] the issue will be the state of your mdadm array, not lack of cryptsetup [20:43] vlt from normal boot https://i.imgur.com/D7ebyoM.jpg [20:45] tomreyn there shouldn't be a raid array, it's only a single drive laptop. this happened after an update, affects all installed kernels [20:45] intok: I’m sorry, I can’t view screenshots. Can you describe the gist of the output? [20:46] vlt mdstat stat lists personalities of linear, multipath, raids 0, 1, 6, 5, 4 and 10, in that order. cmdline says permission denied [20:46] intok: make that "cat /proc/cmdline" on the last line [20:48] intok: Yeah, sorry, I meant *cat*. [20:49] vlt tomreyn ok, bootimage=vmlinz-4.18.0-20-generic root=/dev/mapper/ubuntu--mate--vg-root ro  acpi=noirq pci=noacpi ivrs_ioapic[4]=00:14.0 ivrs_ioapic[5]=00:00.2 [20:50] intok: are you using raid? [20:50] vlt that last bit is required for the laptop to even boot as per Acer's stupid bios, see here https://community.acer.com/en/discussion/comment/619839/#Comment_619839 tomreyn [20:50] EriC^^ no [20:50] intok: did you recently change anything in the bios regarding ahci/raid etc? [20:50] intok: `blkid`, please. [20:51] this may be a non-existing resume device [20:51] EriC^^ see the link to the acer forum above, it's required to boot linux at all on this laptop, had been working since January when I installed 18.10 [20:53] exit [20:53] eix [20:53] t [20:53] exit [20:55] intok: i see [20:55] better [20:55] ? [20:55] intok: try typing this in the initramfs prompt: RESUME= [20:56] then just press ctrl-d [20:56] tomreyn: it's not the swap, it's root fs, [20:56] intok: can you boot a live usb? that would help troubleshooting since you'd have internet access and can upload pastes easily, if tomreyn 's suggestion doesnt owrk [20:56] *work [20:56] vlt EriC^^ tomreyn https://pastebin.com/mkpYXfDR [20:56] TJ-: hmm right. wont work [20:56] intok: Are you currently at the initramfs shell ? [20:57] I've dealt with this issue before [20:57] TJ- see th paste [20:58] intok: which one? [20:58] TJ- yes, booting from internal at the moment, what should I try before I try the live? [20:58] intok: OK, I can talk you through the investigation to determine if you can manually do what is required, or need to do a live-boot recovery [20:59] intok: first check there's a cryptroot config file: "ls -l /conf/conf.d/" [20:59] intok: Try `cryptsetup luksOpen /dev/sda3 a3`. [21:00] hey, is /dev/null not available in initramfs? [21:00] intok: if that file is missing it tells us the root-fs /etc/crypttab is incorrect, or /etc/fstab is [21:00] I'm getting can't create /dev/null: nonexiststen directory on my init-premount script [21:01] j0h: what's the entire command causing that? [21:01] TJ- : if ping -c 1 example.com > /dev/null 2>&1; then [21:01] example.com just example [21:01] but it's an ip in real [21:02] this is what I'm trying to do basically: https://github.com/ceremcem/unlock-luks-partition/blob/master/reverse-tunnel-setup.md TJ- [21:02] root delay might not be long enough? missing modules cat /proc/modules; ls /dev /dev/mapper/ubuntu--mate--vg-root doesn't exist vlt TJ- tomreyn EriC^^ [21:02] the issue is that I am getting ping: not found, sleep: not found as well [21:02] this might be because I'm not using absolute routes? [21:03] there is an issue in there with the creator of that repo with no solution [21:03] TJ- file in question is in step 3: link-with-server.sh [21:03] j0h: look at /usr/share/initramfs-tools/init which is the /init script; you'll see mount -t devtmpfs /dev happens early on [21:03] intok: Is that output of the cryptsetup command? [21:04] intok: try typing 'ls -l /conf/conf.d' as TJ- suggested [21:04] vlt sh: cryptsetup: not found [21:05] TJ-, oh yea, I can see that it's there [21:05] intok: Remember my first guess? :D [21:05] intok: rootdelay isn't the issue, and for everyone: "mdadm: No arrays found..." is expected because the code is looping and sleeping repeatedly waiting for the root file-system to appear; each loop it rescans all devices [21:06] TJ- do you think I might have to copy ping and sleep bins over in hooks (step 2) [21:06] ? [21:07] -rw-r--r-- 1 14 splash -rw-r--r-- 1 43 zz-resume-auto vlt TJ- EriC^^ tomreyn [21:08] intok: so, no cryptroot then, that explains it. [21:08] intok: in which case it is likely cryptsetup is missing too. try "which cryptsetup" [21:09] TJ- " which cryptsetup " gives no output [21:10] just starts a new prompt [21:11] j0h: a little hint; the Debian/Ubuntu dropbear-initramfs package already does all the needed initramfs hooks to get itself installed AND generate the key [21:12] intok: OK, so somehow you've broken the root-fs config in fstab and/or crypttab, and this when the initrd.img was generated the tools were not included. [21:12] TJ- ok, so this is all completely foreign territory to me. [21:13] intok: So, first thing to do is reboot, interrupt at the GRUB stage by tapping Esc key, then choose the Advanced sub-menu and from that boot an older kernel version (because its initrd.img should not have been changed if you're lucky). If trying all the older kernels doesn't work you'll need to recover using a live-ISO [21:13] TJ-, I'm trying to setup a reverse ssh tunnel, hence the script needed [21:14] I fixed the missing ping and sleep issue, now it's just throwing that noexistent /dev/null error [21:14] j0h: Oh, reversing it, sorry, I get what you mean [21:15] pragmaticenigma: FYI! Everyting I asked for help for earlier today is now fixed and in order. Thanks to the people helping here. Just so you know - all is good. [21:19] TJ- initramfs should include sleep right? [21:19] j0h: not sure if this is related but... busybox-initramfs binary doesn't contain the 'ping' applet [21:20] ahh TJ-, I'm using dropbear-initramfs [21:23] j0h: right.... but that only provides the ssh client. busybox provides the shell and core tools and busybox-initramfs does *not* have a "ping" so your code trying to call ping is going to fail, run this to see the list of included applets: "/usr/lib/initramfs-tools/bin/busybox" [21:24] ooh TJ-, you're right [21:24] it doesn't list ping [21:24] humm [21:24] j0h: to prove it do "sudo lsinitramfs /boot/initrd.img-$(uname -r) | grep ping" [21:25] yup, no ping [21:25] vlt TJ- tomreyn EriC^^ tried all installed kernels, no dice [21:26] vlt TJ- tomreyn EriC^^ go live iso? [21:26] i guess you'll have to then [21:26] intok: Yes, follow what I described first. [21:27] TJ-, do you know why it's causing an error with /dev/null being nonexistent directory? [21:27] because the init file does show it sets up /dev [21:30] j0h: probably cause 'ping' is missing the line is being mis-interpretted.. anyhow let's fix tthe missing ping. In step 2 you'll see "copy_exec /usr/bin/dbclient /bin" - add after that "copy_exec /bin/ping /bin" [21:31] TJ- done, will update and try again [21:31] j0h: That *should* help the ping issue and may as a side-effect prevent the complaint about /dev/null if it originates from the check_internet() function [21:31] gotcha [21:31] will see what happens now, thanks [21:35] oh [21:35] stuck on boot loop now TJ- [21:35] =x [21:35] j0h: so ping works :) [21:36] ooh, damn, I think the interface doesn't get setup yet TJ- [21:36] so it will never really be able to get a ping [21:38] j0h: are you booting with "break=premount" so you can manually investigate ? [21:38] j0h: are the correct network driver kernel modules added into the initrd.img, and any required firmware? [21:39] TJ-, the network does work, eventually [21:39] j0h: in initramfs? [21:39] as I was always able to connect to it locally [21:39] yea [21:39] so it worked fine when I ssh'd in from local (without the tunnel) [21:39] j0h: then you need to determine if it is still coming up but those scripts are failing [21:40] j0h: so "break=premount" will let you explore [21:40] ahh, so is there a way to break free from the bootloop and get into to set break=premount? [21:40] like, I'm not sure how I can even uncrypt anymore with kvm [21:41] doesn't even wait for the prompt, just boots straight away [21:44] TJ- even recovery more doesn't work [21:44] j0h: "break=premount" is added to the kernel command-line, so presuming you're using GRUB, interrupt the boot by tapping Esc to edit the entry [21:45] i'm back [21:46] ahh TJ-, I see it now [21:48] TJ- which line should I edit, one that starts with linux or initrd? [21:48] and do I add break=premount at the end of either of those lines? === intok is now known as intok-linux-work [21:51] (03:29:43 PM) vlt: intok: `CRYPTSETUP=Y mkinitramfs ...` from a USB booted chroot environment should fix this temporarily. === jstein_ is now known as jstein [21:52] j0h: line beginning linux and add towards the end of the line; it's just another kernel option [21:53] hi [21:53] I am just setting up a virtual server (droplet) on Digital Ocean [21:53] Now I have to choose the OS and version [21:53] nice TJ-, thanks [21:53] The latest LTE is 18.04 [21:53] TJ-, think I'm in intramfs prompt now [21:53] so, 19.10 will be the next LTS of Ubuntu [21:54] j0h: so find out if the network is available [21:54] now I am not so sure, if I should really pick 18.04 [21:54] should I go with 19.04 ? [21:54] but 19.04 is not LTE [21:54] LTS [21:54] it says netwokr is unrechable [21:54] when i try to ping localhost [21:54] Rojola: LTS are every 2 even years, 16.04, 18.04, 20.04 [21:54] j0h: so you need to check first if the interface is up, and if it has an IP address, and then if there's a route etc [21:55] TJ-, oh, my bad, I thought it was 1,5 years [21:55] TJ-, well, what should I do then? 18.04 LTS or 19.04 ? Or, (offtopic) Debian 9.7 [21:56] vlt https://pastebin.com/Ayc8ZZie am I supposed to be running this from the terminal? [21:56] I personally like Debian and Ubuntu equally much [21:56] Rojola: I'd always go for an LTS on a server [21:56] TJ-, gotcha, I've used ip link show and it does show the ethernet interface (eno1) but no ip info [21:56] TJ-, ok - and if I had to choose between LTS and Debian 9.7 (9.9 is the latest version of Debian) [21:56] er the terminal against the USB iso? [21:56] should I go with Ubuntu? [21:56] TJ-, I need to catch the last train home but will take this with me and continue to debug when I get home, thanks for all your help so far [21:57] j0h: ip link only shows the Ethernet state; for IP address use "ip addr show" and for routes "ip route show" [21:57] I need to install Docker, Sandstorm, etc. [21:57] ip route show is empty [21:58] it's weird that this isn't getting auto ip [21:58] j0h: is dhclient running for that interface? [21:59] dhclient says eno1 link is not ready [21:59] intok-linux: I’m not sure what your question is. [21:59] thenthen I got an error saying shown not gound [21:59] vlt should it be directed at sda1? [22:00] sbin/dhclient-script TJ- [22:00] chown* [22:01] j0h: does "ip link" show the interface is up or down? [22:01] ip link shows lo 1: lo and 2: eno1 but no indication if it's up or down [22:02] j0h: og of course it's the busybox cut down version [22:02] intok-linux: Maybe there’s an easier way now but I would chroot into the root fs (with `mount -o bind /dev dev` prior) and run it there. [22:02] TJ-, how come when I just did a normal busyboxy/dropbear ssh setup, the network was setup and connected fine [22:03] but this script to create the reverse ssh is having issues with the network? [22:03] like, what made it not work? [22:03] j0h: actually it should still show status, I ran that busybox here and for my wifi link get "2: wlp2s0: mtu 1500 qdisc mq qlen 1000" [22:03] ahh TJ- I see it [22:03] it shows UP for eno1 [22:03] sorry, missed that === jje_ is now known as jje [22:04] need to dash, sorry TJ-, I'll be back online to debug in an hour [22:05] j0h: if the ethernet link is up it'd look like this "2: enp1s0: mtu 1500 qdisc mq master br-lan state UP mode DEFAULT group default qlen 1000" [22:05] oh dear. someone suggest i run jdupes on my music. i ran it in prompt-to-delete mode and it seems i have around 8500 or more duplicates. dunno if i want to go thru each one and say yes [22:05] j0h: note my example is a port attached to a bridge so it also shows "master br-lan" [22:05] TJ- ping works now for some reason [22:05] I don't know if running dhclient enabled it? [22:06] j0h: presumably it took some time for dhclient running in the background to onbtain a lease [22:06] yea [22:06] j0h: YES! [22:06] that's dhclient's job [22:06] it threw some chmod unrecofnized option error but must have worked somehow [22:06] I thought it failed [22:17] vlt TJ- EriC^^ tomreyn from the install: not seeing the proper grub config file, just the sample pastes from the install /etc/fstab https://pastebin.com/MzGtFtxS /ect/cryptmount/cmtab https://pastebin.com/02VfTYst [22:19] vlt EriC^^ tomreyn /i/ really have to go at the moment, can you please continue any new ideas on the ubuntu forum post? https://ubuntuforums.org/showthread.php?t=2421770 === jje_ is now known as jje [22:31] so the "cryptmount" package is involved there, i never knew this existed. and it's got an empty configuration file, which probably makes things break. i'm used to /etc/crypttab only. whether or not this exists on this system and what its contents may be we do not know. [22:31] and intok is gone. [22:39] What is the approved method for specifying a DNS server, or just resetting everything to DHCP, in Ubuntu server 18.04. [22:40] I just uninstalled PiHole from my server to move it to its own dedicated device, but there's a setting somewhere telling the server that it is still its own DNS server, so it fails to connect to anything external. [22:41] check /etc/resolv.conf [22:44] I edited that earlier and got it to resolve, but it doesn't persist after reboot. [22:48] I ran into a problem [22:49] so, I've successfully set up this amazing software called "wekan" [22:49] now I must import a project which already exists [22:49] unfortunately, the software does not let me upload the .json file [22:49] I have to paste the .json file [22:50] So, I opened the .json file, selected everything, and copied it [22:50] then I wanted to paste it, but nothing happened [22:50] I guess, it's too much textr [22:50] *text [22:50] the json file (pure text) is 6 MB large [22:50] can you pipe it with terminal commands? [22:51] Gallomimia, not sure... the software (wekan) runs in a browser [22:51] ewwww [22:51] nah it's beautiful [22:51] you can't paste it in stages? [22:51] www.wekan.io [22:51] yes well that could work.... but one tiny mistake and everything is messed up [22:51] well [22:51] don't make a mistake [22:52] kworker 50% CPU/libreoffice 100% CPU https://imgur.com/a/enE3Uvb [22:52] isn't there a clipboard manager who can handle it? [22:52] you'll have to forgive me for being prejudiced against some "software" that's trapped inside a browser container instead of running natively [22:52] SaaS is against my religion. all hail Richard Stallman [22:52] Gallomimia, I respect your opinion, but politely beg to differ [22:53] Gallomimia, I personally appreciate portable software accessable from various devices with the same storage behind [22:53] However, I understand that there are many, many use-cases for natively running software [22:53] well. okay. but the blog post he wrote about specifically cautions against using it because you'll run into problems like this. [22:53] so yes, I absolutely get your point [22:54] i think one day we'll get the best of both worlds [22:54] Rojola1: use the browser's dev console to read the file and put the text in the input [22:54] Gallomimia, how would I share tasks with a team of people without the cloud? [22:54] where your computing environment will be inside a cloud of sorts, portable everywhere, but yet still owned by you [22:54] deltab, I know the dev console, but everything else is unclear. Please be so kind and elaborate. [22:54] by owning and operating your own server [22:54] Gallomimia, I am operating my own server. [22:54] Gallomimia, Wekan is open source [22:55] Rojola1: you need the text in a textarea, presumably [22:55] well then. maybe you can plant that file in the server files? [22:55] deltab, let me quickly verify that it's really a textarea - brb [22:56] deltab, yes, it is a textarea: [22:56] document.querySelector('textarea.js0import-json').value = 'somehow get the json here'; [22:57] I could perform an ajax request [22:57] sounds like there's a couple of tricks you can do to skip the browser aspect in this operation. i assume it's a one-time setup process? [22:57] maybe curl can do it for you [22:57] Rojola1: yep [22:57] Gallomimia, 30 boards [22:57] and to get the json text, you could start a webserver (python -m SimpleHTTPServer) [22:58] every board has it's own json file [22:58] deltab, I am using Apache anywa [22:58] +y [22:58] alternatively, is it just a POST? [22:58] let me check. [22:58] brb [22:59] deltab, https://dpaste.de/hg0o/raw [22:59] shouldn't these json files go somewhere in the server directories? do they really need to be imported thru the web interface? [22:59] it's a weird
without an action and without a method [22:59] send a request with the network panel open, select the request, choose 'Copy as curl command', paste and edit the curl command to read from the file [23:00] Gallomimia, they go into the MongoDB [23:00] deltab, I don't think this is a post-form [23:00] can you import it into the database then [23:00] it just says [23:00] Gallomimia, absolutely no clue how they store it [23:00] and its not even MySQL [23:00] i mean.... it's starting to look less and less beautiful if you're unable to get it to cooperate [23:01] I have zero experience with MongoDB [23:01] good time to learn [23:01] form with no action? probably relies on js [23:01] I will try the things you mentioned [23:01] deltab, Gallomimia - THANK YOU! [23:01] I will give you feedback soon [23:01] check what happens in the network panel when you submit it [23:02] ok [23:02] I must minimize the chat window for a moment to solve it [23:02] brb [23:02] (few minutes) [23:02] Gallomimia: MongoDB is very easy to learn. It's just JSON. [23:03] or use the api? https://wekan.github.io/api/ [23:09] if mongoDB is "just" JSON it sounds like he should still be able to drop the files directly into the database storage area [23:09] or possibly process them with some script [23:19] * Gallomimia_ kicks his ghost [23:20] :) === Gallomimia_ is now known as Gallomimia [23:21] you didn't see anything. [23:22] TJ-, back online again, do you know how I could start dropbear again in initramfs shell? [23:23] so that I can get back into the main ubuntu server [23:25] j0h: start the dropbear SSH server? [23:25] yea [23:25] I have an IP now after running dhclient [23:25] j0h: I'd have thought it'd have started by now [23:25] humm [23:25] maybe it did, let me check if I'm connecting to the right ip/port [23:26] if you stopped at the premount stage, I'd have thought dropbear would have been started earlier in order to receive the LUKS commands [23:26] check if it is running with "ps | grep dropbear" [23:27] how can I split 1 big textfile in 20 textfiles equally sized? [23:27] without adding line breaks or spaces [23:27] a json-file to be exact [23:27] everything else has failed, so I want to manually copy and paste it into the textarea [23:28] TJ- dropbear is running but not with the options I had set [23:28] it's almost like it's using standard config [23:28] Rojola1: see "man split" [23:28] as I defined another ssh port that's not 22 [23:28] Rojola1: the split command [23:29] thank you [23:29] j0h: I guess it depends on how and where you defined that setting [23:29] TJ- in /etc/dropbear-initramfs/config [23:30] but that's fine, I can connect on 22 [23:30] but it's doing a weird timeout on askpass [23:30] before it was prompting to enter decrypt password [23:31] j0h: what did you add to ../config ? DROPBEAR_OPTIONS="-p 1234" [23:32] yea, and also -s -j -k -I 60 [23:32] j0h: maybe there's something in those reverse-ssh scripts altering things? [23:32] can I disable the script from running at boot? [23:32] so I can boot back and revert the line I added which I think cause this to get stuck in this state? [23:33] the bin copy for ping [23:34] that reverse-ssh script is broken in so many ways as it seems [23:34] j0h: I don't see how that is going to break it; I'd expect something deeper is going on. ping isn't going to affect SSH or anything else [23:34] it's the reboot that's affecting it [23:34] so ping works, but because the interface isn't coming back up fast enough, it goes to a boot cycle [23:35] I don't know well enough about initramfs or anything involved to try and fix that script :( [23:35] j0h: right, which shouldn't happen since that has a 60 second watchdog sleep... if it takes 60 seconds for the connection to get an IP address there's something else wrong [23:35] oh you're right [23:36] j0h: did you just install the parts I saw in that link or other scripts from that same repository (I saw others listed but didn't check what they were) [23:36] just that link [23:36] j0h: better give it me again I'll take a swift look, but it's very late here I need to get to bed [23:36] I used this to setup dropbear-initramfs: https://hamy.io/post/0009/how-to-install-luks-encrypted-ubuntu-18.04.x-server-and-enable-remote-unlocking/ [23:37] oh I understand, past midnight here too :( [23:37] https://github.com/ceremcem/unlock-luks-partition/blob/master/reverse-tunnel-setup.md [23:37] so I did the hamy tutorial to do the crypt unlock stuff which worked as expected [23:37] then the reverse tunnel stuff which broke it all [23:39] how can I delete a snap package without traces? [23:39] I messed up the snap "wekan" installation [23:39] j0h: a quick note; is the reason you're not operating dropbear on port 22 to avoid the ssh client "scary warning" about different keys when SSHing to the real root-fs after unlock? [23:39] I would like to purge it including the database [23:39] TJ-, I just changed it from 22 as a force of habit [23:40] less people trying to hit and brute standard ssh port [23:40] (yea security by obscurity is not really much but at least it'll get hammered less) [23:41] j0h: OK, well, if you want to avoid the ssh client warning just do this: https://paste.ubuntu.com/p/cWkc9XZkcS/ [23:42] ahh ok TJ-, will do [23:44] j0h: I presume in check_internet() you've altered the hard-coded "example.com" in the ping? [23:44] yea, I've added my own IP [23:45] j0h: that really should be $LINK_UP_SERVER [23:45] I updated that on my script to use the VAR [23:45] instead of duplicating [23:45] also I may have messed up the script when I renamed the sleep bin [23:46] I used /sbin/sleep and when I ran initramfs shell, /bin/sh scriipts/link-with-server.sh [23:46] it complained about /sbin/sleep not being a thing [23:46] maybe that's what's causing the issue? [23:47] hum, is there an edtior in the initramfs shell to edit that script and maybe rename the sleep to /bin/sleep [23:48] j0h: I don't see any call to /bin/sleep there only "sleep" (you'd not usually get an error messages /adding/ to the command's path in an error message)! [23:48] I added it in my script when it complained about no sleep being a thing [23:48] but I added /sbin/sleep [23:48] which when I run the script in the shell says doesn't exist [23:49] j0h: the busybox shell sohuld be proving /bin/sleep [23:49] /sbin/sleep is not a thing I don't think, it should have been /bin/sleep [23:49] yea, this was before we fixed the ping issue [23:49] j0h: it tells me here it is in the initrd.img [23:49] what is referring to /sbin/sleep then? [23:50] so you see on that script it's just sleep, well I did something stupid and replaced all the "sleep" occurrences to "/sbin/sleep" [23:50] j0h: oh, you mean you've edited the script so it has absolute paths in it? [23:50] and you added /sbin/sleep by mistake? [23:50] yup [23:50] bingo :D [23:51] j0h: oh, that's easy to fix [23:51] woo [23:51] j0h: create a sym-link: "ln -s /bin/sleep /sbin/sleep" [23:52] j0h: also, something to think about, if that script reaches the watchdog() function and then causes a reboot, the script has twice got "online" back from those calls to check_internet() [23:53] ahh [23:53] so, that ln fixed the script when running in the shell, at least it printed watchdog console logs etc [23:53] thanks for that [23:56] j0h: can you find out where dhclient is being called from, and find out why it takes so long to get an IP address. Best way to do that is monitor the DHCP server itself if you can, watching for the REQUEST [23:57] ok TJ- will do [23:57] j0h: because I do NOT see a call to dhclient to get an IPv4 [23:57] the weird thing is, when I did the first tutorial, it got an IP and everything was fine [23:57] so it _must_ get called somehow [23:58] but when you bring that reverse ssh script into the mix, something is messing it up [23:58] j0h: try (on a booted system) "grep -rn dhclient /usr/share/initramfs-tools/" and you will see that the only call to it is for IPv6 (dhclient -6) [23:58] j0h: so unless you get a result showing a call to dhclient -4, or just dhclient (without -4 or -6) then there is your problem [23:59] j0h: in which case your script needs to start dhclient -4 ... [23:59] TJ-, any idea how it worked without the reverse ssh stuff? [23:59] cause that's just weird