[00:00] TJ-, I first noticed it, because I was wondering why I'd get prompted for my pw twice [00:01] I got confused when light-locker wanted it's password (which it doesn't have) not mine! [00:01] When it gets silly now I Ctrl+Alt+F1, "pkill light-locker", Alt+F7 back to Desktop [00:01] I need some help :( I'm stuck in "emergency mode" I think it had something to do with my video drivers. I finally managed to get it into a tty mode with nomodeset, and I can startx successfully under maintainence mode [00:02] I was stuck without the ability to call up tty with ctrl+alt+f [00:02] just a black screen [00:02] even after apt-get remove --purge nvidia* from rw mounted recovery [00:03] trying systemctl default was no dice either. [00:03] just brings me right back to emergency mode. [00:03] Term1nal, which kind of nvidia device do you have? [00:03] jvelasquez: GTX 1070m [00:04] Term1nal, and which version of driver did you try? and which source was it, from distro, or from nvidia.com ? [00:04] graphics-drivers/ppa [00:04] Term1nal: best thing to begin with is to get the system booting to CLI only, but in multi-user mode. At the boot menu edit the entry and add to the kernel-command line (linux ...) "systemd.unit=multi-user.target" and press Ctrl+X to boot with that. [00:04] tried nvidia-current, tried nvidia-367 [00:05] tried reloading nouveau [00:05] Term1nal, did you blacklist nouvuo or whatever it is? [00:05] Term1nal: with full multi-user you've then got all the ttys and you can manually test starting the GUI with "sudo systemctl start lightdm.service" (or whatever DM you're using) [00:05] jvelasquez: at this point trying to get rid of nvidia since it seems to be problematic [00:05] I agree with TJ- [00:05] but I will try that TJ- [00:05] one moment [00:06] TJ-, doesn't he also need nomodeset in that kernel cmnd line? [00:06] the only thing that even got me this far was removing gfxmode=$linux-gfxmode or whatever the line was. [00:06] and adding nomodeset [00:06] and adding nouveau to modules [00:07] TJ-, would new video drivers get into his initramfs ? [00:08] TJ-: do I just append it to the same line within the quotes? [00:08] TJ-, he describes his problem as sounding early in boot [00:08] jvelasquez: Term1nal no, not loaded until rootfs is available [00:09] jvelasquez: "nomodeset" may help if you get a black screen when starting with the multi-user.target, yes [00:09] TJ-, so, if it's on the network, sshd running, he could ssh right into the machine? [00:09] jvelasquez: it'll use the standard VESA modes then [00:09] jvelasquez: indeed. [00:09] trying it now with the added parameter [00:09] Term1nal: don't need the quotes - I use those to show commands are separate from my commentary [00:09] Term1nal, if you have a second machine, turn on sshd as soon as you can too [00:10] I still have the line: gfxmode=$linux_gfx_mode [00:10] Term1nal, well get rid of it [00:12] jvelasquez: no dice. hits emergency mode once again [00:13] reached target system time synchronized is the last systemd output [00:14] there is one failure, which is the mounting of a storage filesystem on a secondary hard drive, non-critical to operation [00:14] TJ-, so, in single usr mode, if he disables X, then he'd have no issue reaching a degraded multi-user mode right? [00:15] "failed to mount /media/Storage.... " which gives dependency failed for Local File Systems [00:15] jvelasquez: it'd be a regular multi-user mode [00:15] Term1nal: that's not an issue, external media [00:15] Term1nal, try, systemd.debug-shell=true and then trying to get to tty9 after boot [00:16] Term1nal: emergency mode means something else is wrong. [00:16] I "appear" to be in a tty currently. [00:16] since it goes straight from grub to the sytsemd [ OK ] message scrolls [00:16] Term1nal: you can always add "init=/bin/bash" to the kernel command line and then you've got a shell in rootfs, where you can disable things without systemd running [00:17] yeah, I can ctrl+alt+f and get to other TTYs, but they're all just a cursor [00:17] chmod a-x `which Xorg` would work unless you run wayland [00:17] I appear to be on the first tty (ctrl+alt+f1) [00:17] I'll try the debug-shell thing [00:18] but it's a horrible way to do it. [00:18] do I still want to keep the multi-user.target in the parameters as well? [00:18] i think so. [00:19] you want to try, multi-user, and open a debug shell [00:19] but the debug shell doesn't always work. but init=/bin/bash will almost always work. but you won't have process control here, and so you'll have to be Very careful what you type. [00:20] debug shell did not appear to work. [00:20] still hit emergency mode. [00:21] ok. init=/bin/bash [00:21] but you won't have process control. [00:21] you'll want to think about what you want to do. which is why I had thought, chmod a-x `which Xorg` [00:21] jvelasquez: can't I just use the root shell from "maintainence" ? [00:21] I can do package management and so forth there [00:21] Term1nal, yep [00:22] okay, so I'm at the root shell [00:23] well, emergency mode root shell, I put in the root password to get it [00:23] if you've got systemd, you can just do "systemctl disable lightdm.service" to prevent the DM and GUI from starting [00:23] get a consensus here on a plan of action first [00:24] TJ-: strangely, I can actually startx from this shell and get an almost-working gui [00:24] just not seem to be able to launch apps [00:24] it seems most people run lightdm in ubuntu, so TJ's suggesting sounds best. [00:24] disable lightdm [00:25] ok, done [00:26] TJ, what about, systemctl set-default multi-user ? [00:27] Term1nal, that should be good, reboot [00:27] sorry, I'm in and out. Fixing an android phone [00:28] TJ, what about grub's gfxmode ? what's with that? [00:28] TJ, he didn't have a problem before installing this nvidia package, yet now grub's gfxmode is enough to disable the screen? [00:29] if you've got editing control now, then edit "/etc/default/grub" and uncomment the line "GRUB_TERMINAL=console" and then "update-grub" [00:29] that way the system will remain in a text-only console [00:29] get some stability back then you can focus on getting the correct drivers sorted [00:29] somehow, update-grub isn't changing things. [00:30] error? [00:30] no, just doesn't do anything. [00:30] Term1nal, i think that's good [00:30] Term1nal: have you "mount -o remount,rw /" ? [00:30] Term1nal, grub-install is the one that has output [00:30] when I boot and check the config (press e) it still shows a paramter that shouldn't be there anymore because I removed it from /etc/default/grub [00:30] ahh [00:30] update-grub should report the entries it adds to /boot/grub/grbu.cfg [00:31] I had added novdeau.modeset=0 [00:31] noveau* [00:31] but I remove it and update-grub, it persists [00:31] but I know for a fact I'm in rw [00:32] Term1nal: it's not in "/etc/default/grub" ? [00:32] TJ-: definitely not there. [00:33] Term1nal: is /boot/ on a separate file-system? is it mounted currently at /boot/ or are you generating the file into the rootfs, which will then have the real /boot/ FS mounted over it [00:33] which would mean the file you're updating neve gets seen by GRUB :) [00:33] Term1nal: "grep boot /et/fstab" ... anything? [00:33] Term1nal: "grep boot /etc/fstab" ... anything? [00:34] TJ-: if I run update-grub from "recovery mode" proper it works [00:34] but not when I hit "emergency mode" [00:35] Term1nal, also, i believe update-grub likes to keep old configs as backups [00:35] i imagine emergency mode might be a onionfs for rootfs or something [00:36] well, scrap that it didn't actually change it [00:36] still has that noveau thing I added in before [00:36] I checked /boot/grub/grub.cfg and it shows the entries as they should be. [00:37] Term1nal: do you have multiple distros installed on the PC? [00:38] Term1nal: I'm wondering if GRUB's root isn't the one you think it is [00:38] EFI or BIOS? === capella|away is now known as capellaGoesBOO [00:39] i swear, sometimes EFI has two different part for boot, /boot and /boot/foo [00:40] Term1nal, what about, update-grub -o /tmp/myGrub [00:40] does that work? [00:42] yeqah [00:43] do you know if you're using EFI or BIOS? [00:43] efi [00:44] can you confirm if /boot is mounted ? [00:44] yes it is. [00:44] rw ? [00:44] "grep boot /proc/mounts" [00:44] ahh! [00:44] always check /proc/mounts ! [00:45] not `mount` [00:45] I got it to stay this time [00:45] mount lies ~ [00:45] but I had to go in at boot and remove the gfxmode line again [00:45] so it wouldn't just show a black screen [00:45] mount used to consult /etc/mtab which could be stale; now it's usually a symlink to /proc/mounts so should be correct, but reading the kernel's list directly avoids any doubt [00:47] /proc/mounts shows /boot/efi [00:47] /dev/nvme0n1p1 [00:48] how do I find out what is "causing" it to hit emergency mode? [00:49] Term1nal: "journalctl -xb" may help [00:49] no idea. journalctl -xe ? tail -1000 /var/log/syslog ? [00:53] what should I look for? [00:55] unit apparmor.service has begun starting up. [00:56] followed immediately by something abut /var/log duplicate line for path [00:56] then reached target emergency mode [00:57] Term1nal: have you attempted booting with "systemd.unit=multi-user.target" and did it still hit the emergency.target ? [00:57] systemd-tmpfiles-setup.service has begun starting up, then next line is started emergency shell [00:58] TJ-: I have that systemd.unit=multi-user.target at the end of the line containing "linux..." [00:58] it's in the grub.cfg as well [01:00] Term1nal: OK, so it's something beforehand, for sure. [01:01] do you have a separate file-system for /tmp/ ? [01:01] The cause of the emergency may appear in messages earlier in the log - it is running several service start-ups in parallel [01:03] TJ-: no, /tmp is not a separate filesystem [01:03] the only error I have in red text is regarding that secondary storage drive. [01:04] local-fs.target: triggering onfailure dependencies [01:05] Term1nal: that must be it then; Systemd must have somehow generated a dependency for it [01:06] Term1nal: check in /var/run/systemd/generator/ see if you can see one there that matches the external media. Removing it might be enough [01:06] hmm [01:06] could be right. [01:07] so just get rid of the .mount entry for that storage drive? [01:08] there's an entry -.mount [01:09] I can't cat it or anything [01:10] cause it keeps trying to see the dash as a paramter [01:10] parameter* [01:10] in that situation do ./-.mount [01:10] (or a full path) [01:11] I guess it's the root .mount [01:11] because it says Where:/ [01:11] What:/dev/mapper/vgroup-lvroot [01:12] so I removed the symlink in local-fs.service.requires/ [01:12] to the storage drive [01:14] Anyone know much about apparmor ? my /etc/hosts file is a symlink to /etc/@/$(hostname)/etc/hosts, and so now I get tons of errors about apparmor. how can I get apparmor to follow the symlink? [01:18] jvelasquez: you can set additional permissions under /etc/apparmor.d/abstractions/ [01:19] hello ppl, I'm attempting to write a script/plugin for gimp that just scales selection by certain criteria .. can you please gimme output from followung command on hw displays ? returns thing like this: http://sprunge.us/GTGV ,, command is: [01:19] perl -e 'for $x ("-v", "--listmonitors", "-q") { print foreach("### xrandr $x\n", qx/xrandr $x/); }' | curl -F 'sprunge=<-' http://sprunge.us [01:19] thanks much!! === patche is now known as _sebbot [01:21] Anyone that uses preload..that sees it actually increases speed? [01:22] ntz curl: (6) Could not resolve host: sprunge.us [01:24] I was looking at the list of startup applications,and I was curious as to what 'has' to be there for ubuntu to startup well....what can i remove so startup is faster, any ideas? [01:24] ArMedic: in an age of SSDs high speed RAM I don't think it'll gain much in observable speed IMHO. Unless you are using a platter based disk for your Ubuntu installation I can't see it doing much [01:24] ntz, ohh. it was me@! [01:26] ntz, you need this? http://sprunge.us/CZDj [01:26] Actionparsnip, I don't have a SSD so I am guessing this one is platter based. [01:27] ArMedic: bum used to be able to enable / disable startup services and you can research what each is. I'm not sure if it works with systemd. Maybe there is something similar for systemd [01:27] jvelasquez: :D, thanks much!! ,, groovy [01:28] ArMedic: if you move your Web browser cache to tempfs it can make Web browsing a little faster. [01:29] Actionparsnip, This computer has 2.16GHZ cpu, and 8GB RAM...I just believe the reaction time for opening stuff should be faster than what it is. Trying to improve on speed. [01:31] Actionparsnip, Its a fresh install so that isn't the issue. [01:31] ArMedic:seems systemctl can be used to disable and enable startup services. See man pages. Your system isn't too bad. I've ran Ubuntu on way less. If you switch to XFCE or LXQT then you will get a snappier desktop [01:32] ArMedic:reducing vm.swappiness will keep applications in RAM. Set it to 2 and it should help [01:33] Actionparsnip, Thanks! [01:33] ArMedic:nw [01:38] please help me i think i'm going crazy: $ ls -al bin/ioncube_encoder55_8.3 >>> -rwxrwxrwx 1 moshe moshe 830784 Oct 1 00:05 bin/ioncube_encoder55_8.3 [01:38] $ bin/ioncube_encoder55_8.3 >>> zsh: no such file or directory: bin/ioncube_encoder55_8.3 [01:39] use a different shell maybe? [01:39] Myrth: is it on a mount that has noexec option? [01:39] # bin/ioncube_encoder55_8.3 >>> bash: bin/ioncube_encoder55_8.3: No such file or directory [01:39] TJ-: how do i check this [01:40] Myrth: the other possibility is that it's for another architecture. You can check arch with "file " [01:40] use stat [01:41] ioncube_encoder55_8.3: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux.so.2, for GNU/Linux 2.2.5, stripped [01:41] Myrth: well, you know where the file is, so work out which mount it is on and check the options for that mount with "cat /proc/mounts" [01:41] it was working today [01:41] Myrth: is the system 32-bit i386 install? [01:41] and just suddenly [01:41] i've tried that for both _64 and not [01:41] bhuh [01:41] you're right [01:41] i'm on 64 [01:41] shouldn't it be able to run 32 [01:42] you need to ensure the foreign arch is added and install multilibs [01:42] how do i do that? [01:42] Myrth: "sudo dpkg --add-architecture i386" [01:42] Myrth: then "sudo apt update" to get the 32-bit package lists fetched [01:43] Myrth: you will also likely need "sudo apt install multiarch-support" [01:44] is there anyway to install a .deb package using aptitude only? [01:45] thank you [01:45] ntz: xrandr: unrecognized option '--listmonitors' [01:45] TJ-: apparently i do have multiarch-support already installed [01:46] did --add-architecture and apt update [01:46] still runs only x64 [01:46] ntz: http://sprunge.us/KGLd you need this? [01:46] Myrth: OK, so you need the libraries for i386 that that program requires. Identify them with "ldd " [01:47] Myrth: you'll see some "not founds" or similar message there. For each, you have to discover which package contains that file and then install it (for the i386 arch)... [01:47] ok i guess will have to ask the publisher of the soft for those [01:48] Myrth: ... so for example lets say it is missing "/usr/lib/foorbar.so" - then first see if the equivalent 64-bit library is installed with "dpkg -S /usr/lib/foorbar.so" ... if you get a package name there you can install the i386 version with "apt install :i386" [01:48] Myrth: rinse and repeat for each missing lbirary. [01:48] thank you so much [01:49] Myrth: if you can't find a 64-bit package installed you either need to install "apt-file" or use http://packages.ubuntu.com/ and advanced search to find packages containing the file, and then do "sudo apt install :i386" [01:49] i guess i'm used too much to windows where both 32 and 64 work seamlessly [01:49] Myrth: they do on Linux if you install from the package archive because the packages declare their dependencies so everything required is installed automatically [01:50] Myrth: but you put something on the system manually, you're responsible for installing its dependencies too - whether that is 64 or 32 bit [01:51] Myrth, they don't make ioncube for 64bit ? [01:51] jvelasquez: they do, but not for all loader versions [01:52] old ones don't have _64 version [01:55] Myrth, Wow64 on windows you mean? [01:56] i don't really have a good understand about windows internals [01:56] *good understanding [01:56] I had thought it consumes a GB+ just for wow64 [01:56] me neither. too complicated. [01:56] and manual is incomplete [02:10] so I have to reboot after changing an apparmor profile huh? [02:18] jvelasquez: I'd have thought "systemctl restart apparmor.service" [02:18] ohh it's a service? [02:19] jvelasquez: although I wouldn't be surprised if it uses inotify to watch it's config directories to automatically detect changes. many services do [02:19] the part that loads the profiles into the kernel is, yes [02:19] same as iptables loading rules into kernel netfilters [02:19] iptables is a service now? [02:20] no, but iptables-restore is often run by firewall services at start-up to reapply the rules [02:21] the point being a userspace 'service' runs once to load profiles into the kernel. They don't need to continue to run unless they monitor their config directories for changes and reload them [02:22] I'm surprized inotify rules aren't part of systemd unit scripts [02:22] systemd uses inotify to watch for changes to its own configs [02:23] echo "inotify IN_CLOSE_WRITE /etc/apache2" >> apache.service [02:24] I'm gonna whine to the systemd guys about that right now [02:25] hi im having a problem with wired connection on a ubuntu machine [02:25] mattfly, me too. [02:25] ive installed wicd but wanst able to connect [02:25] does anyone know how to stop irssi from joining a channel at startup? [02:25] its not a startedup [02:25] its been some hours of googling and trying [02:25] mattfly: what release of ubuntu? [02:26] 14.04 [02:26] it used to work normally but then there was a lightning and i suspect the board can be damaged [02:26] mattfly: so that usually uses ifupdown's /etc/network/interfaces [02:26] but kinda weird since the router is ok [02:26] mattfly: ahhh... now you tell us :) [02:26] how can i know if there is still a board that works there? [02:27] there were other compters connected to the lan and they werent damaged [02:27] mattfly: have you used ethtool to check if the device looks good? also /var/log/kern.log for any reported errors from the driver. [02:27] not really [02:27] mattfly: when the cable is in ethtool should indicate if the link is active. kern.log should report it UP as should "ip link show" [02:27] wont dmesg show anything? [02:27] because it doesnt [02:28] mattfly: maybe try a different cable too [02:28] dmesg is where kern.log gets its messages from [02:28] yeah it is a looong cable [02:28] it exists because wifi didnt work on that distance [02:28] mattfly: the kernel should report the device being detected and initialised. You've got an 'eth0' node for it I assume? [02:29] mattfly: if there's an ethX node then the driver is fine [02:29] there is [02:29] i didnt get that error [02:30] mattfly: ok, so I'd goto "ethtool" and do some diagnostics, check it reports the link active and connected (at some speed) [02:30] and there is a eth0 interface after ifconfig [02:30] okay, how should i use ethtool? [02:30] mattfly: is the PC configured to use DHCP? [02:30] yeah it was [02:30] mattfly: and is it getting an IP at all? [02:30] but i also tried static [02:30] no [02:30] mattfly: OK, focus on the link status first then [02:30] and when i try dhclient [02:31] it starts sending those dhcp idk what now and never ends [02:31] or after i do ifup eth0 [02:31] mattfly: start with "ip link show eth0" ... it should say it's UP [02:32] mattfly: check the system does have ethtool with "which ethtool" ... obviously without net you can't easily install it :) [02:33] mattfly: if it's there do "sudo ethtool eth0" and look for the line "Speed: XXXXMb/s" if it isn't there then the link isn't established [02:34] mattfly: can try "mii-tool eth0" to show link status [02:34] okay [02:34] mattfly: however even when mii-tool indicates link ok, I've had situations where packet loss exceeded 80%, eg in the case of bad cables or fried adapters [02:35] ip link show eth0 says [02:35] mattfly: no-carrier means it thinks the cable is unplugged or broken [02:35] it is not plugged [02:35] i will plug and check [02:35] im back soon [02:35] mattfly: as a test, with a short patch cable, can you put a laptop or other PC next to this problem one and see if they can bring up the link between them? That'd help establish if it's a loooooong cable issue or something in the hardware. Test it each end with "sudo ip link set eth0 up" [02:37] mattfly: I have to go, hope you get it sorted === Piece_Maker is now known as Acou_Bass [02:47] have i missed something? [02:47] even tho i plug the cable it says "eth0: no link" [02:47] as output of mii-tool eth0 [02:47]

mattfly: as a test, with a short patch cable, can you put a laptop or other PC next to this problem one and see if they can bring up the link between them? That'd help establish if it's a loooooong cable issue or something in the hardware. Test it each end with "sudo ip link set eth0 up" [02:47] but the cable works in another computer [02:47] okay [02:48] ill get a cable [02:50] mattfly: how long is "long"? ethernet is rated to go up to 100m if there's no joins or similar disturbances in the cable [02:51] mattfly: so if it works in another computer but that one says no carrier (rather than connecting but giving epic packet loss) then it's likely that the ethernet port is damaged [02:52] no [02:52] its like 50 m [02:52] and the same cable works on my laptop [02:52] i was talking at first over it [02:52] TJ-, it went right onto their todo list. master:TODO: - maybe add support for inotify events [02:53] mattfly: then likely the ethernet port is busted unfortunately [02:53] let me connect both computers [02:54] i just got a wire [02:54] okay both connected [02:54] what do i do [02:55] TJ-, the inotify rules for systemd services, that is. [02:55] TJ-, #systemd peeps like the idea. [02:55] mattfly: set them both up, check mii-tool [02:55] it still shows no link [02:55] serious man :/ [02:55] this lightning? [02:56] a longe wire over the roof [02:56] mattfly: seems that way [02:56] yeah [02:58] hey [02:58] !!! [02:58] on my other computer [02:58] when im connected with this one mii-tool returns: enp7s0: no autonegotiation, 10baseT-HD, link ok [02:58] if i unplug from it it says no link [02:58] this from the other computer [02:58] suspecious [03:00] 10baseT seems rather suspicious on its own [03:00] that's rather old technology [03:00] yes it is old [03:00] old as fuck [03:01] coax-cable kind of old? [03:01] coax cable? [03:01] im using normal lan cables [03:01] ah, sorry [03:01] I'm mixing it up with 10base2 or 10base5 [03:02] mattfly: sounds like your ethernet adapter is indeed toast - the other end can detect that a cable is there but the other end won't talk to it [03:02] aw [03:02] shit [03:02] they could just be failing at autonegotiation [03:03] but the computer that doesnt work says eth0: no link still [03:04] try forcing the same on both sides [03:04] forcing the same what? [03:04] you might also need an actual crossover cable if it's going from computer to computer and you have old hardware [03:05] same ethernet mode [03:05] i am tryign to connect this old on to lan [03:06] ok [03:06] im testing to see if the lan board is toasted or not [03:06] actually it is a laptop [03:06] both are [03:06] so you are connecting two computers together directly or via a switch/hub? [03:07] directly [03:07] just to see if it would work [03:07] hardware would have to be ancient to not have autoneg right? my computers had that in 2005 [03:07] this old laptop is about that age [03:08] mattfly: modern ethernet cards do crossover automatically, ancient ones didn't [03:08] so a direct computer-to-computer connection would require a crossover cable [03:08] the plan is not crossover is just to make sure if the board is gone or not [03:08] is there a way to make screen have the bottom bar in byobu? [03:09] autonegotiation (which you're relying on) relies on the cabling being correct [03:09] so no link if cabling is wrong [03:09] aw jesus, ok [03:09] plug it into a hub or switch [03:09] so what else should i check for? [03:09] if light comes on, all good [03:09] no light, all bad [03:09] i dont have a hub right now [03:10] or home router [03:10] i have a lan cable that comes from the home router that is a bit far away right now [03:10] ok [03:10] this cable works on my laptop, but not on the old one [03:11] i can see the eth0 interface on the old one as output of ifconfig and others like miitool and ip link show eth0 [03:11] but i i plug the cable it remains as no link [03:11] what's at the other ends when you do those tests? [03:12] a home router [03:12] that should be fine [03:12] but i has no lights [03:12] it * [03:17] okay its time to swuitch to a rhaspberry [03:26] hey [03:26] thanks for all good night === r0Oter is now known as r00ter [03:29]