=== lotus|i5 is now known as lotuspsychje | ||
=== [1]MrMobius is now known as MrMobius | ||
Phruis | what is the windows 10 default font | 00:37 |
---|---|---|
doug16k | Segoe UI I think | 00:40 |
doug16k | I don't think ubuntu has that font though | 00:42 |
Phruis | I just downloaded tahcoma | 00:46 |
Phruis | tahoma i mean | 00:46 |
wawowe | Hi, what service is reverting my manually set ipv4 addresses back to nothing after I set them? | 00:49 |
OerHeks | there are some restricted fonts, https://help.ubuntu.com/community/RestrictedFormats/Microsoft_Fonts | 00:50 |
OerHeks | wawowe, i guess you. how do you set ipv4 manually? | 00:50 |
OerHeks | direcly into interfaces file? | 00:50 |
wawowe | OerHeks, ifconfig interface ip | 00:51 |
wawowe | OerHeks, after a minute or so it changes back to being unset | 00:52 |
sarnold | ew | 00:53 |
OerHeks | and what linux is this? | 00:53 |
wawowe | ubuntu | 00:53 |
wawowe | 16.04 | 00:53 |
wawowe | i have done 'service networking stop' but it still happens | 00:54 |
wawowe | connection is reset while trying to ftp over the interface | 00:54 |
sarnold | don't try to stop the networking service; you're just supposed to ifup and ifdown specific interfaces. trying to stop or restart networking will quite often just wedge the system solid, since upstart and systemd communicate via the networking.. | 00:55 |
OerHeks | why not just set a profile in networkmanager? | 00:57 |
=== zenguy is now known as oii | ||
=== oii is now known as zenguy | ||
OerHeks | if networkmanager is running, sure it reverses back to dhcp ( if not set custom networking) | 00:58 |
sarnold | wawowe: does ip addr show confirm that the ip is *removed*? connection reset is usually a different error | 00:58 |
eelstrebor | is there something going wrong with the internet? i'm using dns 1.1.1.1 and it's real slow | 01:00 |
Phruis | eelstrebor, yeah man my cat messed up the internet | 01:00 |
Phruis | eelstrebor, spilled my cup of water all over the cable | 01:01 |
sarnold | my ec2 instance can ping 1.1.1.1 with about 6ms latency | 01:01 |
eelstrebor | thanks for the serious response | 01:01 |
eelstrebor | 1.0.0.1 is intermittent also | 01:02 |
sarnold | of course 1.1.1.1 is going to be anycasted so if my packets went to a different instance than your packets, that doesn't prove much | 01:02 |
OerHeks | lolz cloudflare | 01:02 |
wawowe | OerHeks, networkmanager is disabled | 01:02 |
eelstrebor | Query time: 15 msec or no response at all | 01:03 |
wawowe | I ran ifconfig with 'up' at the end. So far I still have connection | 01:03 |
giaco | whaat would you use to deploy a private chat service on VPS for mobiles? | 01:03 |
OerHeks | some issues, but no real problemshttps://www.cloudflarestatus.com/ | 01:03 |
sarnold | giaco: irc is the only service I remotely know, so that'd be what I would deploy.. | 01:04 |
=== zbenjamin is now known as Guest91928 | ||
=== zbenjamin_ is now known as zbenjamin | ||
giaco | sarnold: not really smartphone ready .... or not? | 01:08 |
* eelstrebor wonders if his isp is doing something to slow things down | 01:09 | |
kenperkins | anyone had success with spotify in 19.04? I keep getting crashes at startup (deb not snap) | 01:09 |
sarnold | giaco: I've heard good things about https://quasseldroid.info/ and https://weechat.org/about/interfaces/ | 01:12 |
OerHeks | we have no spotify deb package, so seek support by the owner/maintainer? | 01:17 |
OerHeks | it is good we are limited to our own issues | 01:17 |
kenperkins | @OerHeks hm. this page begs to differ: https://www.spotify.com/us/download/linux/ | 01:17 |
kenperkins | regardless, the snap worked even though the deb didn't | 01:17 |
OerHeks | so, thanks for bringing up a non issue. | 01:18 |
kenperkins | I was asking about a legit issue, the deb didn't work and I was looking for some guidance. i also tried the other option and it did work, that doesn't invalidate my question | 01:20 |
OerHeks | i think you find No support for that deb package here | 01:21 |
kenperkins | ah, you know what, I think i totally misunderstood you, and as a result, I misspoke | 01:21 |
kenperkins | for some reason, they way you said "we" made me think you meant we, spotify, I'm so sorry. you're right, that's a spotify issue, and my apologies | 01:22 |
OerHeks | you could check the version that deb installs, against the snap 1.1.10.546.ge08ef575-19 | 01:23 |
OerHeks | 189,9 MB .. wowie.. | 01:23 |
huggybear404 | what command to spin up / down drives ? | 02:01 |
The_Letter_M | If I wanted to add the ppa https://launchpad.net/ubuntu/bionic/+package/mongodb how would I do it? | 02:04 |
OerHeks | that is a package, not an PPA | 02:05 |
The_Letter_M | I kinda knew that. I'm trying to get the 32bit mongodb installed on my Raspi2 running hte Ubuntu Server image. I've been looking all over and this is the first repo I've been able to find with the right architecture | 02:10 |
The_Letter_M | How would I add the ppa using the correct info fromt he url? I mean even if it's the whole PPA? | 02:10 |
sarnold | huggybear404: hdparm ought to be able to do it for many drives.. | 02:14 |
sarnold | The_Letter_M: apt install mongodb should do it | 02:15 |
OerHeks | i find no official ppa | 02:15 |
OerHeks | i hope there are arm builds? | 02:16 |
OerHeks | arm64 | 02:16 |
OerHeks | https://launchpad.net/ubuntu/+source/mongodb/1:3.6.3-0ubuntu1.1 | 02:16 |
sarnold | ohhhh I didn't notice that The_Letter_M wanted 32bit builds | 02:17 |
OerHeks | huggybear404, all i remember is: eject && eject -t | 02:19 |
Joik2ww | hello lads | 02:21 |
tonyt | hi | 02:21 |
huggybear404 | eject drives ? nice idea, maybe with some dynamite to help it along ? | 02:21 |
lotuspsychje | huggybear404: volunteers are trying to help you here | 02:22 |
huggybear404 | pity I cant find any info on this, would be nice to have a bit raid with spare drives and be able to boot up the spares from another country | 02:22 |
huggybear404 | so I could restore it even when not home | 02:22 |
huggybear404 | guess that is a very obscure command | 02:23 |
huggybear404 | maybe reserved for special sysadmins that know the secret handshake ? | 02:23 |
Joik2ww | I'm in crappy situation with kali linux :< | 02:24 |
lotuspsychje | you can stop trolling now huggybear404 | 02:24 |
Joik2ww | I've done everything but still can't get GUI running | 02:24 |
OerHeks | raid hotswap or something like that? | 02:24 |
huggybear404 | I also struggle getting reliable drives , the smart seems pretty bad , every drive use diff smart format and the program runs generic | 02:24 |
sarnold | huggybear404: for the enclosures where hdparm doesn't work, this should do the job http://sg.danny.cz/sg/smp_utils.html (in the smp-utils package) | 02:24 |
The_Letter_M | Yeah I was trying to find the armhf arch of MongoDB | 02:25 |
OerHeks | Joik2ww, no support for kali, seek their channels? | 02:25 |
OerHeks | The_Letter_M, maybe on their builds? https://docs.mongodb.com/manual/tutorial/install-mongodb-on-ubuntu/ | 02:25 |
Joik2ww | OerHeks well those don't differ so much so why can't ask... | 02:26 |
huggybear404 | now I regret getting cheap drives, I fear they will soon die and I cant monitor bad smart data, since in their wisdom they use some bits to record errors and other bits in SAME field records sectors written, so the smart tells me my drive is dying with millions bad sectors while happily ignoring the bits that record the bad sectors so I guess I need to rewrite my own smart or make it somehow | 02:26 |
huggybear404 | translate | 02:26 |
huggybear404 | anyone tryed to make smart read actual data decoding the drives in use ? | 02:27 |
The_Letter_M | OerHeks: No. When I add their repo it doesn't have the packages. Mongo says they ONLY provide 64 versions and when I ooked through their repos manually the no longer have 32 arm packages anywhere | 02:28 |
sarnold | The_Letter_M: if you have a big machine you can use to *crosscompile* the package, you could use apt source to download the sources, and modify these Architecture: lines: https://sources.debian.org/src/mongodb/1:3.4.18-2/debian/control/#L39 | 02:28 |
sarnold | The_Letter_M: .. I don't know how to cross-build debian packages but there's got to be a guide somewhere :) | 02:28 |
The_Letter_M | You don't need a heavy duty machine to cross compile | 02:29 |
sarnold | but you probably need a machine that can do more than two gigabytes of address space to link this thing | 02:29 |
OerHeks | i am sure your pi is 64 bit server capable | 02:29 |
sarnold | oh definitely check that first! I don't know pis well so assumed you knew it was 32bit | 02:30 |
The_Letter_M | It's a pi2. It's using a 32bit ARM processor | 02:31 |
The_Letter_M | the pi3 and 4 that are 64bit | 02:31 |
OerHeks | oh right, raspi 2 is the original risc arm | 02:32 |
OerHeks | no, then you are stuck. | 02:34 |
swift110 | oh | 02:38 |
huggybear404 | hm tryed google, cant find any spin command on either hdparm or smp, what is smp supposed to do ? Do I need to build a relay thing to cut power to each drive and remote this ? Was hoping these days drives are smart enuff to have the spin up and down supported. I also wish all drives would use same format so I dont need to make my own smart translater and reprogram every time I order a new | 02:39 |
huggybear404 | drive model | 02:39 |
huggybear404 | would be sweet to have smart send email alert when a drive start redirecting sectors so I know when it starts fail before it damage any data | 02:40 |
huggybear404 | dreaming of a box that tells me what drives are good and if one start wear out, it emails "replace drive xx" and I could just send a command to have a spare drive take over then when im home again yank the bad drive and put a new in and it automatically takes over. Maybe this kinda thing is just for google datacenters using special os ? | 02:42 |
OerHeks | i think it is controller dependent, maybe the #ubuntu-server guys or #debian have a clue | 02:44 |
huggybear404 | pity its soo hard to build a server that protects its data and tells me what drive or data gets corrupted , seems i end up with tons of desperate backups and I have no way of knowing which version of a file is corrupted. I dream of a server that checks each file and if one goes bad it should tell me what drive is going bad and restore good file automatic | 02:45 |
huggybear404 | I need to to all this in the controller ? seems areca is one of the most used ? | 02:46 |
huggybear404 | have some but I never figured out how to use their special software and drivers | 02:46 |
huggybear404 | I try ask in server | 02:46 |
=== akem__ is now known as akem | ||
JimBuntu | huggybear404, we all dream of this, but outside of a system that runs full auditing to know when a user has interacted with a file, the best it could do is alert you to changes... unless the checksum/hash changes regularly without any changes to the modified on timestamp, then it can alert you to possible corruption. | 03:07 |
huggybear404 | yes im thinking sha1 or something checksums for all files and maybe a monthly scan to see if any file changed that should not have, but seems lot of work to make it work and automate | 03:32 |
huggybear404 | and where to put all the checksums ? maybe a a separate folder tree that mirror the real data is best ? | 03:33 |
huggybear404 | anyone made any headway doing this any easyer ? | 03:33 |
huggybear404 | takes a while to play terabytes of mp3 f.ex to see if any got bad | 03:33 |
ryuo | huggybear404: there's already technology for that. it's called ZFS, though it won't help much against intentional modifications. | 03:36 |
ryuo | i guess snapshots could help there. | 03:37 |
huggybear404 | yes I read about jounaling filesystems but seems hard to make them alert me if a drive or file goes bad | 03:37 |
huggybear404 | most of the data is media that should never change so I was thinking make most of the volumes read only to the network | 03:38 |
ryuo | huggybear404: ... ZFS isn't your regular filesystem... | 03:39 |
ryuo | it has features most filesystems don't. | 03:39 |
ryuo | everything has checksums and can have redundant copies made for a chance to fix data corruption. | 03:39 |
ryuo | all automatically. | 03:39 |
ryuo | it comes with a form of RAID as part of the package. | 03:40 |
ryuo | and i suppose also LVM too. | 03:40 |
ryuo | incidently mdadm has no such guarantees. | 03:41 |
ryuo | huggybear404: but what you're describing is already standard parts of ZFS. | 03:41 |
ryuo | zfs scrub? checks all data blocks for integrity. | 03:42 |
ryuo | can be run as frequently as desired. | 03:42 |
ryuo | now, the drawback is ZFS requires a fair amount of RAM for reasonable results. | 03:43 |
ryuo | I run my ZFS server with 16G of ECC RAM. | 03:44 |
JimBuntu | ZFS is really cool and fills a lot of gaps, but I didn't think it was for the faint of hear or commoner yet. I guess I'm outdated | 03:44 |
ryuo | funny, i thought you were describing BTRFS. | 03:44 |
JimBuntu | lolz, def not for everyone | 03:45 |
ryuo | i hear far more often about BTRFS data loss than ZFS data loss. | 03:45 |
JimBuntu | I would imagine so | 03:45 |
ryuo | i'd just use BTRFS if i felt i could use it without a lot of headaches. | 03:45 |
JimBuntu | ryuo, is ZFS in the normal repos? I only saw zfs-FUSE | 03:46 |
ryuo | where have you been? it's been part of the standard ubuntu kernel since 16.04 | 03:46 |
huggybear404 | but will zfs scrub tell me if any file is bad ? | 03:46 |
JimBuntu | Ah, I'm on 16.04, thanks ryuo | 03:47 |
ryuo | huggybear404: yes, at least as far as on disk silent corruption goes. | 03:47 |
ryuo | huggybear404: it can't help with intentional data writes. | 03:47 |
ryuo | incidently i almost never get inconsistencies in my scrubs. | 03:48 |
ryuo | seems to be rare, but it can still happen. | 03:48 |
huggybear404 | hm I remember years ago I ran zfs and was impressed by it, seemed to stay intact even when fuses blew out the power | 03:48 |
huggybear404 | so sfs and raid is the way to go ? Im a bit scared of raid, if I mess up order of drives or more than 2 drives die then its all dead | 03:49 |
ryuo | order of drives isn't that important, honestly. it's mostly how they're grouped. | 03:49 |
JimBuntu | with a single drive, if 1 dies it's all gone | 03:50 |
SwedeMike | huggybear404: all modern raid have superblocks with unique identifiers per drive, so it'll figure out the correct order itself. | 03:50 |
huggybear404 | but a redundant server maybe with no raid should help, but need to automate backups when files get added etc | 03:50 |
ryuo | huggybear404: well, ZFS helps a lot. it can do stuff that your conventional LVM + MDADM + filesystem layers cannot do. | 03:50 |
huggybear404 | so If I mess up the order of the plugs it finds out itself ? | 03:50 |
SwedeMike | huggybear404: correct. | 03:51 |
ryuo | yes. i run a mirror ZFS at home. it works fine. | 03:51 |
huggybear404 | nice, but what is modern ? mdadm does this ? | 03:51 |
SwedeMike | huggybear404: both md-raid and zfs has superblocks and will figure out what drive needs to go where. | 03:51 |
huggybear404 | i have an old tired server I want replace now, need a new with faster bigger drives and I want even better data protection | 03:51 |
ryuo | MDADM is software RAID but is not very sophisticated. It can't do much advanced stuff because it's not integrated with the filesystem layer. | 03:52 |
ryuo | ZFS is probably your best solution then if you got the hardware for it. | 03:52 |
huggybear404 | is there any goot program to scan for similar files ? | 03:52 |
ryuo | uh, that could be handled by ZFS deduplication at the block level. | 03:52 |
ryuo | but that's VERY ram hungry. | 03:52 |
huggybear404 | would save tons time when comparing old backups to see what files are missing on the new drive and what files have changed | 03:53 |
ryuo | ... actually | 03:53 |
ryuo | i think you might want to read about ZFS send/receive commands. | 03:53 |
ryuo | that's one of the uses for iirc | 03:53 |
huggybear404 | zfs need special hardware ? | 03:53 |
ryuo | can be used for incremental backups. | 03:53 |
ryuo | No, not really. It works with any 64 bit platform with enough RAM. | 03:54 |
ryuo | ECC is recommended but not required. | 03:54 |
huggybear404 | so it only stores whats changed ? thats nice, dont want overwrite what already is there correctly | 03:54 |
ryuo | well | 03:54 |
ryuo | it can sync an external ZFS pool or w/e. | 03:55 |
ryuo | sending only what's changed. | 03:55 |
ryuo | i figured that's what you meant? | 03:55 |
ryuo | otherwise i guess you could try mounting snapshots at different points and comparing the files there. | 03:55 |
ryuo | but if this is something you want, you'll likely need to use snapshots. | 03:55 |
ryuo | this lets you keep old copies of data around even if deleted, up to the snapshot point anyway. | 03:56 |
huggybear404 | hm that sounds promising, ill try read about these more | 03:56 |
huggybear404 | yes I want make sure even if a virus try wipe it all then it doesnt kill the backup | 03:57 |
huggybear404 | so if a file changes it should not overwite but store a new version with date | 03:57 |
ryuo | and it also replaces the main use cases of LVM by providing its own filesystem setup.. it allocates from the overall pool. | 03:58 |
huggybear404 | whats lwn ? | 03:58 |
ryuo | Logical Volume Management. | 03:58 |
huggybear404 | what should I start with to get this going ? seems theres years of research just to learn about this ? | 03:58 |
ryuo | not really? there's guides for installing to ZFS on root. | 03:59 |
ryuo | almost everything is in man zfs or man zpool | 03:59 |
ryuo | there's also #zfsonlinux for more serious questions. | 03:59 |
swift110 | hwy all | 04:00 |
huggybear404 | yes install zfs seems easy back then I just selected use it but to make it report bad files etc need a lot scipting and learning commands | 04:00 |
ryuo | huggybear404: not that much? setup a cron or systemd timer job to run it periodically and put the report somewhere, email or something. | 04:00 |
huggybear404 | would be realy cool to have a nice hotswap rack as well, when a drive goes bad just yank it and plug in a new one and its done | 04:00 |
huggybear404 | but seems hard, last I tryed replace a drive I was unable to make it accept the new drive coz the serial number didnt match i guess | 04:01 |
huggybear404 | should be muche easyer to tell the system that drive died, now use the new one instead | 04:02 |
huggybear404 | maybe this only exist in the billion $ datacenter systems ? | 04:02 |
ryuo | well.. | 04:02 |
ryuo | hotswap isn't always a priority.. depends how much you can afford some downtime to make the exchange. | 04:03 |
ryuo | plus having to identify which drive is the problem from the exterior may be a challenge. | 04:03 |
OerHeks | not all sata controllers are hot swapable, with a "devices/pci0000:00/0000:00/ataX/hostY/scsi_host/hostY/scan something" | 04:04 |
huggybear404 | yes thats the main thing | 04:04 |
huggybear404 | not yank the wrong drive and kill the while rack forever | 04:04 |
ryuo | well, if it worries you so much, you can use a whole lot of disks with a more fault tolerant RAID mode. | 04:04 |
huggybear404 | wish I had some nice leds to show status, like green for ok idle, yeallow working , and red for dead-replace | 04:05 |
huggybear404 | better than raid 6 ? | 04:05 |
huggybear404 | it takes 2 spare right ? | 04:05 |
ryuo | Well. | 04:05 |
ryuo | There's always RAID1 if you're ultra paranoid. | 04:05 |
ryuo | It can survive the most drive failures. | 04:05 |
huggybear404 | well I plan a full redundant backup server anyway so more like raid 6 + raid 1 | 04:06 |
ryuo | otherwise ZFS offers ZRAID1-3 or RAID10 emulation. | 04:06 |
huggybear404 | pity good areca controllers are soo pricy | 04:06 |
ryuo | who needs hardware raid when you have zfs ;) | 04:07 |
huggybear404 | so mortgage the house to get a new one ? or try make it work with an old sata2 board ? hm | 04:07 |
ryuo | I run ZFS in a cheap ass proliant. | 04:07 |
ryuo | the older gen8 proliant ml10 v2 | 04:07 |
huggybear404 | speed isnt really that critical so the old will prob. still do the job | 04:07 |
ryuo | I upgraded mine from the original RAM and put in a quad core CPU. | 04:08 |
huggybear404 | is there an easy way to get the red leds for bad drives ? | 04:08 |
huggybear404 | maybe try build my own but it will need programs to run it as well I guess | 04:09 |
ryuo | Maybe, but i've only heard of that for rack mount servers. | 04:09 |
huggybear404 | maybe an arduino board and wired to the mdadm program ? | 04:09 |
ryuo | one option? note the serial #s of your drives with your own labels. | 04:09 |
ryuo | that would help a lot. | 04:09 |
huggybear404 | if drive failes turn on red light, blaring siren horn and email sysadmin ;=) | 04:10 |
ryuo | don't overthink this if you don't actually need so much sophistication. drive failure isn't usually an everyday event unless you have a lot of servers. | 04:10 |
huggybear404 | why need 16 gb for zfs ? how I know how much I need ? | 04:10 |
ryuo | Well. I'd ask in #zfsonlinux. | 04:11 |
huggybear404 | I tested with new server and 4 gb and it did run out of ram once so guess I might need more | 04:11 |
ryuo | 8G is what I originally used. | 04:11 |
huggybear404 | big drive ? | 04:11 |
OerHeks | one needs 4gb+1gb, for all functions like blocktransfer | 04:11 |
ryuo | but i put in another 8G so i could do more. | 04:11 |
ryuo | i sometimes build software. | 04:11 |
ryuo | that can get pretty RAM hungry too. | 04:11 |
huggybear404 | is it to keep the fat ? | 04:11 |
ryuo | huh? | 04:14 |
=== jmcgnh_ is now known as jmcgnh | ||
TheSov | is there any way to install ubuntu server without partitions? | 04:43 |
ryuo | What? It's not possible to boot a system without a partition table. How could you install without partitions? | 04:44 |
elias_a | TheSov: Do you mean installing on _one_ partition? | 04:50 |
TheSov | elias_a, no i mean on no partitions at all | 05:57 |
TheSov | I tried a differnt way i put a partition table on a small disk, and put /boot on that disk and im trying to put / on a partitionless disk but the ubuntu installer wont proceed | 05:59 |
ryuo | TheSov: ... why? | 06:00 |
naribia | hello, I'm formatting an external harddrive using Disks, it shows a spinning activity icon next to the drive, so im guessing it's busy, but i have no other information, how do I know if its busy or something else is wrong? | 06:01 |
ryuo | partitionless disks are possible but, why? | 06:01 |
ryuo | the amount of overhead of a partition table is neglible in this day and age. | 06:01 |
TheSov | because partitions are prevent ease of expansion when using a GPT disk | 06:01 |
ryuo | uh... expansion? you mean changing their size later? | 06:02 |
TheSov | I am fully automating the deployment of VM's | 06:02 |
TheSov | the size of the disk is unknown at this time | 06:02 |
ryuo | Ok... then use LVM. | 06:02 |
TheSov | no | 06:02 |
ryuo | that's what it was created for. | 06:02 |
TheSov | lvm has a max block size limit | 06:02 |
ryuo | so does ext4. | 06:03 |
TheSov | correction max size block transfer | 06:03 |
TheSov | of 512k | 06:03 |
TheSov | direct ext4 does not | 06:03 |
ryuo | and why does that matter? | 06:03 |
TheSov | it interferes with the "iops" limiter of openstack | 06:04 |
ryuo | well, what's stopping you from just making a single partition using the entire space? | 06:05 |
TheSov | so my preference would be to have a disk with no partition on it at all so that on boot the ext4 can be resized online | 06:05 |
TheSov | because when you expand a gpt disk, there is no utility on ubuntu right now that can be scripted to do it without manual intervention | 06:05 |
ryuo | uh... why would it need to be expanded? it's already at maximum capacity. | 06:05 |
TheSov | fdisk cant do it at all, and parted freaks out and asks you to repair the disk | 06:06 |
TheSov | no it isnt | 06:06 |
TheSov | the disk is size is not known | 06:06 |
OerHeks | why did livepatch not install this kernel vuln? | 06:06 |
OerHeks | https://usn.ubuntu.com/4144-1/ | 06:06 |
OerHeks | what is actually the use of livepatch? | 06:06 |
ryuo | TheSov: uh... sure? but if the disk has a partition table allocated for its size, that size never gets any bigger. | 06:06 |
ryuo | TheSov: so, why would you need to expand it? | 06:06 |
TheSov | this is for a deployment image | 06:07 |
ryuo | oh. | 06:07 |
TheSov | since the image size is finite it doesnt have the true size of the disk | 06:07 |
ryuo | so you'd what? dd the image to the drive? | 06:07 |
TheSov | the image is like 8 gigs but the deployed systems could have TB's of disk space | 06:07 |
ryuo | how do you separate them then? | 06:07 |
TheSov | their is a boot script that regens keys and such | 06:08 |
TheSov | only once | 06:08 |
TheSov | it created a "i finished this already file" and never does it again | 06:08 |
ryuo | how do you know where each begins and ends? | 06:08 |
TheSov | eh? | 06:08 |
ryuo | unless they just exist as files in the host filesystem? | 06:08 |
TheSov | the images are deployed to a ceph RBD | 06:08 |
TheSov | which is individual to each system | 06:08 |
ryuo | so what is each VM fed then? a block device? a disk image file? | 06:10 |
TheSov | im other words the RBD is allocated, the image is DD'd on that RBD and then KVM kicks off the VM using the RBD as its primary disk, that disk could literally be any size | 06:10 |
ryuo | it sounded untenable from what i understand. | 06:10 |
TheSov | so i need to vm once run to expand its disk to the size of the rbd at first boot | 06:11 |
ryuo | what is RBD even... | 06:11 |
TheSov | its like an iscsi target | 06:11 |
ryuo | and this is your alternative to LVM? | 06:11 |
ryuo | eh. anyway. | 06:12 |
TheSov | hmmm how to describe.... imagine i could tell a manger system i want a /dev/sdX device of this size and it carves it out of a san and maps it to the local system | 06:12 |
ryuo | so where is the host's root filesystem then? | 06:13 |
TheSov | on its own disk | 06:13 |
TheSov | the RBD is basically a SAN its not local to the server it just behaves that way | 06:13 |
ryuo | so what's the issue? it sounds like it should install normally. | 06:13 |
ryuo | i assumed the host would get a regular setup with this fancy stuff reserved for guests. | 06:14 |
TheSov | the issue is that the VM has a "disk" in it, and when its deployed on the RBD its disk size is whatever the image inside says it is | 06:14 |
ryuo | so how is this related to installing ubuntu...? as a guest? | 06:14 |
TheSov | yes | 06:14 |
TheSov | so i can create the image | 06:14 |
TheSov | oh i see the confusion | 06:15 |
TheSov | yes im installing ubuntu as a guest as a template for the image im building | 06:15 |
ryuo | ... yes, i assumed you were talking of installing to the host. | 06:15 |
TheSov | i would to install it without a partition | 06:15 |
ryuo | sounds doable, but i don't think it fits inside the guided installer. | 06:15 |
=== pabed1 is now known as pabed | ||
ryuo | you may have to use anothr method. | 06:16 |
TheSov | ahh i see | 06:16 |
TheSov | bummer | 06:16 |
ryuo | same way you setup zfs on root i presume... | 06:16 |
ryuo | see debootstrap. | 06:16 |
ryuo | with a bit of effort you can do a manual install. | 06:16 |
ryuo | it should support almost any scenario that the system will tolerate. | 06:17 |
ryuo | not to mention debootstrap is more amenable to automation... | 06:17 |
TheSov | thank you I will look into it! | 06:19 |
tomreyn | TheSov: https://unix.stackexchange.com/a/5168 | 06:19 |
TheSov | worry not tomreyn i know that KVM will boot a disk without a partition, but the ubuntu installer wont cooperate | 06:21 |
tomreyn | but current grub is really too large to fit, which means you need a separate biosgrub storage already then, and embedding it is and always has been 'a hack' | 06:21 |
tomreyn | TheSov: kvm booting in bios mode then, i guess, which i assume will be replaced by kvm + omvf sooner or later, which then probably reintroduces the issue | 06:22 |
TheSov | by then I will come up with something else i doubt KVM will remove that functionality anytime soon | 06:23 |
tomreyn | probably not, it's just that expectatrions against a public cloud may be "uefi" in the future. but maybe this is not public at all. | 06:24 |
TheSov | it is not | 06:24 |
tomreyn | lucky you ;) | 06:25 |
TheSov | internal openstack environment | 06:26 |
TheSov | well we also use ubuntu server bare metal for our ceph storage | 06:28 |
tomreyn | /join #ubuntu-server | 06:29 |
=== jje_ is now known as jje | ||
Industrial | Hi | 07:16 |
Industrial | I have a dell laptop. I have loaded and am using the i915 driver. | 07:16 |
Industrial | Can I use the nvidia driver instead? | 07:16 |
Industrial | I want to use Vulkan technology to play a game using Lutris | 07:16 |
Industrial | I have no clue if the i915 driver supports this :) | 07:17 |
Industrial | Intel Corporation UHD Graphics 620 (rev 07) | 07:18 |
Industrial | `sudo ubuntu-drivers list` rturns nothing | 07:20 |
pantato | how does one apply a patch properly from apt-src? i'm only seeing instruction on patching for the git version | 07:21 |
Industrial | https://www.intel.com/content/www/us/en/support/articles/000005524/graphics-drivers.html | 07:23 |
Industrial | oh Vulkan is supported on 620 ,nice :-) | 07:23 |
mgedmin | pantato: mkdir ~/src/apt-sources && cd $_ && apt source somepackage && cd somepackage-* && patch -p1 < your.patch && dpkg-source --commit | 07:23 |
mgedmin | pantato: dpkg-source will ask a bunch of questions interactively and create a debian/patches/<name-of-patch>, add it to debian/patches/series, and open an editor for you to write some metadata | 07:24 |
mgedmin | pantato: then you can build the patched package with debuild | 07:25 |
pantato | thank you | 07:26 |
Industrial | ssh-irc | 07:27 |
mgedmin | pantato: btw instead of apt source you can use pull-lp-source somepackage distrocodename, then you don't need actual deb-src lines in your /etc/apt/sources.list | 07:28 |
mgedmin | (the downside is you can't use apt build-dep without deb-src lines, which is _very annoying_) | 07:28 |
pantato | erm | 07:39 |
pantato | mgedmin: not sure what this means: dpkg-source: info: 'dpkg-source --commit' is not supported by the source format '1.0' | 07:41 |
mgedmin | ah! | 07:43 |
tomreyn | Industrial: not sure that's obvious, but: the i915 driver / kernel module is for intel graphics (those embedded into the CPU) whereas the nvidia driver / module is for (separate) nvidia chipsets. | 07:44 |
mgedmin | so, my earlier instructions were missing a step: before a debuild do a dch -i to create a new entry in debian/changelog where you can describe your changes and update the version number to something like <upstream-ver>-<ubuntu-ver>-<your-suffix> | 07:44 |
pantato | i also wanna clarify that i'm not trying to commit the patch anywhere, i just want to compile it for my local system | 07:45 |
mgedmin | as for source format 1.0, well, this means this particular package doesn't support debian/patches/ etc. and so just skip the dpkg-source --commit step and do a build | 07:45 |
pantato | ok | 07:45 |
pantato | https://passthroughpo.st/patch-kernel-debian/ i think this guide is the one that i needed, i just needed to change a few more of the trust checks to =n | 07:46 |
tomreyn | Industrial: "ubuntu-drivers list" only enables you to install proprietary drivers (where needed / supported), such as nvidia's. open source drivers, which can also work well, will just work out of the box. | 07:46 |
mgedmin | pantato: the debian new maintainers guide is a good source on debian source package formats, if you're curious and have time: https://www.debian.org/doc/manuals/maint-guide | 07:47 |
ws2k3 | im trying to upgrade mysql 5.6 to mysql 5.7 on ubuntu 14.04. since the mysql repo packages are broken im trying to upgrade it using .deb files. but im not entirely sure what im doing wrong. https://pastebin.com/YyFDSw7s | 08:02 |
lotuspsychje | ws2k3: how many times do we have to tell you 14.04 is eol? | 08:06 |
ws2k3 | lotuspsychje: howmany times i already told you this doesnt change anything. does the eol cause apt to not work propperly anymore? | 08:07 |
OerHeks | ws2k3, indeed, no packages, no updates, no support | 08:07 |
OerHeks | LoLz | 08:07 |
ws2k3 | not true. i can run apt-get update;apt-get upgrade just fine works perfectly | 08:08 |
mgedmin | https://help.ubuntu.com/community/EOLUpgrades | 08:08 |
ws2k3 | OerHeks: but this is upgrading using .deb files not apt-get update;apt-get upgrade | 08:08 |
OerHeks | ws2k3, join ##linux for that, no support here, you seem to have issues we do not fix\ | 08:09 |
OerHeks | lolz | 08:09 |
ws2k3 | i dont need you to fix it. but advise never hurts | 08:09 |
ws2k3 | ubuntu 14.04 is ubuntu right? | 08:10 |
OerHeks | no, 'was' a supported ubuntu | 08:10 |
ws2k3 | OerHeks: wrong. its still ubuntu. it will allways be ubuntu. that its not a supported ubuntu ur right on that. but thats not relevant | 08:11 |
OerHeks | past tense | 08:11 |
ws2k3 | supported or not doesnt change anything about how 14.04 works. | 08:12 |
tomreyn | ws2k3: the ubuntu versions supported on this irc channel are listedon the channel topic. please don't ask for support for unsupported past releases here. | 08:12 |
ws2k3 | tomreyn: k. anyone else got ideas? | 08:14 |
tomreyn | ws2k3: you're still asking on the same channel | 08:15 |
talin | hello. i want to put set a network interface into promiscuous mode, permanently. e.g. "ifconfig eth0 promiscuous". what config file can i set this in? | 08:17 |
tomreyn | talin: hi, which ubuntu version is this? | 08:17 |
talin | tomreyn: 18.04 | 08:19 |
tomreyn | talin: and is this a desktop or server? | 08:21 |
mgedmin | (the reason for these quesions is because we're trying to determine if your network is managed by ifupdown, systemd-networkd, or NetworkManager) | 08:22 |
doug16k | seems ubuntu isn't auto-mounting an ext4 formatted usb flash drive automatically in /media/$USER/something. is that expected? | 08:23 |
talin | tomreyn: this is a server | 08:23 |
tomreyn | talin: by default, 18.04 LTS would use persistent network interface names (you have apparently reverted to the previous naming scheme), and system-networkd managed through netplan on a server. the dedicated server support channel is #ubuntu-server | 08:24 |
doug16k | I had to explicitly `mount` it | 08:24 |
talin | tomreyn: oops, sorry. the real name is enp59s0f1 | 08:25 |
mgedmin | doug16k: hmm, automounting should be possible | 08:27 |
mgedmin | doug16k: do you see the usb flash drive in nautilus? does it mount if you click on it? | 08:28 |
doug16k | ya I figured it would, because external USB hard drive enclosures do it | 08:28 |
tomreyn | talin: https://askubuntu.com/a/1037955 discusses an approach which may work, using a hook script, but this post is from more than a year ago and i don't know whether that is the 'right' way to do it in systemd-networkd land | 08:29 |
doug16k | it's not appearing in nautilus | 08:29 |
tomreyn | talin: i suggest you re-ask in #ubuntu-server, maybe repeated your question there another day if you can't get help today. | 08:31 |
mgedmin | surely creating scripts in /usr/lib/networkd-dispatcher is wrong, sysadmin-created stuff should go in /etc/ ... | 08:31 |
mgedmin | talin: https://wiki.archlinux.org/index.php/Network_configuration#Promiscuous_mode might be helpful! | 08:32 |
mgedmin | ignore that it's from the ArchLinux wiki, systemd-networkd is the same on Ubuntu | 08:32 |
tomreyn | hmm nice suggestion there mgedmin | 08:33 |
mgedmin | huh, this one doesn't even rely on networkd | 08:33 |
Geras | Hi all. Which browser is good for ubuntu that i can configure i2p on? except clearnet firefox and tor browser | 08:34 |
doug16k | mgedmin, ah, nevermind. I think this flash drive is dying. suddenly it does pop up and I get "an operation is already pending" alert dialog | 08:34 |
doug16k | s/pop up/appear in nautilus/ | 08:35 |
Geras | anyone know? :) | 08:36 |
mgedmin | doug16k: any errors in dmesg? | 08:36 |
tomreyn | Geras: any web browser which can use http and https proxies should techically work | 08:36 |
* mgedmin has no idea what i2p is | 08:36 | |
Geras | any recommended one except clearnet firefox adn tor browser? | 08:37 |
lyr | Hi all | 08:38 |
OerHeks | chromium browser perhaps? | 08:39 |
Geras | what about this Palemoon? | 08:39 |
OerHeks | there is opera, not in our repos.. maybe edge-for-linux | 08:39 |
OerHeks | what about this Palemoon?\ | 08:39 |
Geras | yes | 08:39 |
lyr | I've a weird and madening issue : hiting "Ctrl" keep zooming in in Firefox and Tilix (a quake like terminal). I had to lock Firefox zoom (min zoom / max zoom = 100% in about:configs, and switch to gnome terminal which doesn't seems to have this issue. Any troubleshooting way or fix idea is welcome | 08:39 |
OerHeks | you tell us? | 08:39 |
Geras | im asking you | 08:40 |
OerHeks | what about this Palemoon? is not a complete question i can understand | 08:40 |
doug16k | mgedmin, repeated variations of "JBD2: Failed to read block at offset 2245" and "JBD2: IO error -5 recovering block 2245 in log" | 08:40 |
Geras | lol. im talking about the browser for i2p configuration | 08:40 |
Geras | if palemoon is good for that | 08:41 |
mgedmin | lyr: ctrl+mouse wheel is for zooming, and I had unexpected zooms due to mouse whell inertia ("kinetic scrolling") when I pressed Ctrl too soon after I stopped scrolling | 08:41 |
mgedmin | doug16k: yeah ouch | 08:41 |
OerHeks | !find palemoon | 08:42 |
ubottu | Found: W:, W:, W:, W:, W:, W:, W:, W:, W:, W: (and 381 others) http://packages.ubuntu.com/search?keywords=palemoon&searchon=names&suite=bionic§ion=all | 08:42 |
OerHeks | seems like palemoon is not in our repos | 08:42 |
Geras | so i guess chromium then | 08:43 |
lyr | mgdemin: I ruled out this one (removed mouse) | 08:43 |
lyr | as well as behing very careful typing like my grandma to avoid hitting anything unrequired | 08:44 |
doug16k | sandisk.boycott=1, kingston.fanboy=1 | 08:46 |
mgedmin | lyr: do you see any weird events in xev? | 08:46 |
lyr | Tons of ButtonRelease event | 08:47 |
lyr | Googlin' to find which key seems involved | 08:48 |
lyr | Weird, I've no keycode | 08:50 |
lyr | mgedmin: do you understand this output ? https://gist.github.com/rgarrigue/1b68320428f7d69916482816554dd0d2 | 08:51 |
ghost2911 | hello. somebody knows how to set env variables on ubuntu 12 in that way ? systemctl set-environment MYSQLD_OPTS="--skip-grant-tables | 08:53 |
ws2k3 | turn out the package from ubuntu 14.04 WHICH IS EOL is not the same package as the package from mysql repo . so all i needed to do is remove mysql entirely. and then dpkg -i *.deb so its all working now. thanks for the great help. oh wait. | 08:54 |
mgedmin | lyr button 6 is used for mouse wheel scrolling | 08:55 |
mgedmin | lyr: there are pairs of ButtonPress/ButtonRelease events | 08:55 |
mgedmin | lyr: find the bluetooth mouse you have fallen down behind the desk with the wheel wedged into the scroll-left position ;) | 08:56 |
lyr | haha | 08:56 |
mgedmin | buttons 4/5 are wheel scroll up/down, 6/7 are wheel scroll left/right | 08:56 |
mgedmin | sadly xev doesn't tell which mouse device sent the event | 08:56 |
lyr | for my culture, how / where do you get this info ? | 08:56 |
mgedmin | xinput can list all mouse devices known to the system | 08:56 |
mgedmin | I've been a linux user since 1998? I've read a lot of blog posts/howtos/etc.? I remember when mouse wheels were a new invention | 08:58 |
mgedmin | button 6, wow, I didn't know firefox supported ctrl-zooming with the horizontal wheel too | 08:58 |
mgedmin | but it explains why you don't see pages scrolling incessantly -- they're already at the left edge and there's nowhere to scroll more | 08:58 |
lyr | oooh'kay, I've been using linux since 2009 and never had to fight that much with desktop & x11 (cowardly ran is the right word tbh) | 09:00 |
lyr | hell, now it stopped ?... | 09:00 |
Industrial | ssh-irc | 09:00 |
lyr | and my mouse don't have an horizontal scrolling | 09:00 |
lyr | guess I'll further debug it when it start running amok again | 09:01 |
lyr | thanks a lot mgedmin :-) | 09:04 |
zamba | how can i limit the number of logins over ssh for a specific user? | 10:26 |
ducasse | zamba: https://superuser.com/questions/188529/restrict-number-of-ssh-connections-by-user-name-or-ip | 10:39 |
tomreyn | combine this with a Match section | 10:43 |
silv3r_m00n | hi, i am using ffmpeg to remove logo from a video, i remember there was an option that would play the video in a popup video as it is processed, what was that option ? | 11:07 |
zmagii | sup guys | 11:09 |
tomreyn | silv3r_m00n: i *think* there's an ffmpeg channel around here, maybe ask there instead. i'm not used to ffmpeg doing play back, though. | 11:10 |
zamba | ducasse: no, maxsessions is not the same | 11:11 |
zamba | ducasse: maxsessions just states how many sessions a single tcp stream can have | 11:11 |
tomreyn | silv3r_m00n: oh there's ffplay for playback, so i take this part back | 11:12 |
zamba | ducasse: meaning, how many different sessions (sftp +) you can have on top of a single ssh session | 11:12 |
zamba | ducasse: i've seen stuff about limits.conf, but that doesn't stick | 11:12 |
zamba | # ps uaxw | grep <username> | grep sshd | grep -v root | wc -l ; returns 11 hits | 11:13 |
zamba | even though i've set maxsessions to 4 for that user | 11:13 |
zamba | and also i have added <username> - maxlogins 4 to /etc/security/limits.conf | 11:13 |
zamba | /etc/ssh/sshd_config is also set to usepam | 11:13 |
zmagii | silv3r_m00n: i don't see such an option in the manpage for ffmpeg | 11:14 |
silv3r_m00n | zmagii: the delogo thing has a show option which shows the video and the rectangle | 11:14 |
silv3r_m00n | zmagii: https://ffmpeg.org/ffmpeg-filters.html | 11:15 |
silv3r_m00n | but i cant get it to work | 11:15 |
zmagii | so why don't you just crop the output? | 11:17 |
zmagii | https://ffmpeg.org/ffmpeg-filters.html#toc-crop | 11:17 |
=== BrianG61UK_ is now known as BrianG61UK | ||
lisa_ | Hey there! I want to underclock my device using cpufrequtils. | 12:07 |
lisa_ | When I try to set "cpufreq-set -u 2.20 GHz -r" as root it gives me an error. | 12:08 |
lisa_ | any help? | 12:08 |
mceier | I'm guessing - try 2.20GHz without space | 12:10 |
lisa_ | mceier: wow that worked, thanks! | 12:15 |
lisa_ | the obvious thing is sometimes the hardest to see | 12:16 |
lisa_ | weird because the output from cpufreq-info DID have the space... what's the default in unix between floats and unix? | 12:17 |
lisa_ | *units | 12:18 |
Mudchains | Hello, I am migrating ubuntu servers from hyper v to vmware. At one server I am not able to fix boot disk. I deleted the /boot partition in gparted (live cd ubuntu). When I try to do grub-install /dev/sda I am getting a error "failed to get canonical path of /cow" | 12:18 |
mgedmin | lisa_: I suspect cpufreq-set -u "2.20 GHz" -r would also work | 12:19 |
lotuspsychje | come join to #ubuntu-server Mudchains | 12:19 |
mgedmin | command-line arguments are split on spaces, so when you need a space in the middle of an argument, you need to quote it | 12:19 |
Mudchains | thansk lotuspsychje | 12:19 |
lisa_ | makes sense, mgedmin! thanks | 12:19 |
mgedmin | I don't think there's any difference in fixing grub on a server vs fixing grub on a desktop? | 12:20 |
mgedmin | Mudchains: have you booted into a live session? | 12:20 |
Mudchains | mgedmin yes | 12:20 |
mgedmin | Mudchains: I would usually chroot into the actual system before running grub-install | 12:21 |
mgedmin | (you'll also need to mount /proc and /sys, and bind-mount /dev from the live session) | 12:21 |
mgedmin | ah, the chroot is optional if you use grub-install --boot-directory=/mnt/boot: https://help.ubuntu.com/community/Grub2/Installing#via_the_LiveCD_terminal | 12:23 |
Mudchains | mgedmin I tried that (mounting /proc etc, chroot) but then in the chroot session I get the same cow error again | 12:24 |
Mudchains | I will try grub-install --boot-directory=/mnt/boot now | 12:24 |
mgedmin | don't forget to mount the actual (non-live) system on /mnt and its boot partition on /mnt/boot first | 12:25 |
Mudchains | yes | 12:25 |
Mudchains | the grub-install --boot-directory=/mnt/boot gave no error :) I am restarting the server now to check it out | 12:28 |
Mudchains | mgedmin no pxe load screen but a blinking lower dash now :) | 12:30 |
mgedmin | oh hey if you wiped your boot partition then you don't have a valid grub.cfg | 12:31 |
mgedmin | you'll need to boot into a livecd, mount the system and its boot, then chroot and run update-grub to generate one | 12:31 |
BluesKaj | Hey folks | 12:31 |
mgedmin | with proper root filesystem UUID and matching kernel versions | 12:31 |
Mudchains | yes I already fixed the UUID in /etc/fstab | 12:32 |
mgedmin | also, if /boot/ was wiped then you don't have a kernel image, so you'll have to apt-get install --reinstall linux-image-X.Y.Z for whatever version you had | 12:32 |
mgedmin | (which will run update-grub as a side effect) | 12:32 |
Mudchains | the original server is still alive, how can I see the active kernel version? | 12:32 |
mgedmin | uname -r | 12:33 |
Mudchains | thanks :) | 12:34 |
mgedmin | uh, how are you booted into a livecd and how are you mounting its partitions if the server is still aliv? | 12:34 |
mgedmin | that's a recipe for massive data corruption | 12:34 |
Mudchains | the new VM is on new hardware (but network doesnt work) and old vm is still active on old hardware :) | 12:35 |
mgedmin | ah, a copy of the image, of course, VMs | 12:35 |
mgedmin | also you can see the names of installed kernel packages with dpkg -l linux-image* | 12:35 |
Mudchains | mgedmin must I chroot into the boot mount of system mount? | 12:35 |
mgedmin | the ones that have 'ii' in the first two columns are the installed ones | 12:35 |
mgedmin | you need to chroot into the system image so you can apt install stuff in there; the system image needs at least / and /boot and /var and /lib and /usr (but I expect /var and /lib and /usr to be in the same partition as /) | 12:36 |
Mudchains | Thanks mgedmin I will try it out! | 12:39 |
Mudchains | mgedmin I am pulling over a new image of the VM (so /boot is not deleted) ;) and try to fix it with grub-install --boot-directory=/mnt/boot :) keep you posted. | 12:46 |
ace_me | I want a process run faster so nice --15 should be used isn't it ? | 12:50 |
theantz | I just noticed that one of my 16.04 servers more recent installed kernels are linux-image-unsigned instead of linux-image. Is this a bugfix for something? Google found nothing. | 13:00 |
Mudchains | mgedmin still blinking dash unfortunally | 13:01 |
theantz | Except there was a bug sometime last year which required the unsigned package to fix. Should I just switch back? | 13:02 |
flog | 155 | 13:03 |
mgedmin | Mudchains: can you at least get into the grub boot menu? | 13:04 |
Mudchains | mgedmin how can I do that? | 13:04 |
Mudchains | I am now in live cd session to checkout the menu.lst | 13:05 |
mgedmin | uh, menu.lst is grub 1? grub 2 uses grub.cfg | 13:05 |
Mudchains | ah didnt know that | 13:06 |
Mudchains | mgedmin how I enter the menu? | 13:06 |
mgedmin | hold down shift during boot? or press esc at the right moment? | 13:08 |
mgedmin | https://askubuntu.com/questions/16042/how-to-get-to-the-grub-menu-at-boot-time | 13:08 |
Mudchains | mgedmin I see only appaering "grub" with a blinking dash when I hold shift :) | 13:09 |
TJ- | Mudchains: that means that only GRUB's bootstrap code is loaded but it cannot find the core image | 13:11 |
flog | TJ-: i was stupid and the sudoers file was working all along. | 13:13 |
flog | For some reason i thought i could execute the command as root without sudo. | 13:13 |
TJ- | flog: ;s | 13:14 |
Mudchains | TJ- must I install a kernel by chroot then? | 13:16 |
mgedmin | no, apparently grub doesn't get that far | 13:16 |
mgedmin | grub-install failed to install grub correctly somehow | 13:16 |
TJ- | Mudchains: I've not been following along but missing core image means grub-install wasn't run/targeted at the correct boot device, or failed (lack of free sectors/BIOS boot partition ?) | 13:17 |
Mudchains | mgedmin stange, the cmd said everything went succesfull @ grub-install | 13:17 |
TJ- | Mudchains: what command was used? "grub-install /dev/sdX" ? | 13:18 |
Mudchains | TJ- grub-install --boot-directory=/mnt/boot /dev/sda | 13:18 |
TJ- | Mudchains: was that from within a chroot or outside? | 13:19 |
Mudchains | outside | 13:19 |
TJ- | Mudchains: is sda using MBR or GPT ? | 13:19 |
Mudchains | TJ- boot-repair program is complaining about GPT, so I think gpt :) | 13:20 |
TJ- | Mudchains: complaining? in what way? If it is GPT then there needs to be a BIOS Boot partition. | 13:21 |
Mudchains | TJ- boot-repair is complaining there isnt a partion on /dev/sda with a bios boot flag, but gparted says it is.. | 13:22 |
Mudchains | its formatted as ext4 | 13:22 |
Mudchains | TJ- disk layout: /dev/sda = boot disk, /dev/sdb = / , /dev/sdc = swap | 13:23 |
TJ- | Mudchains: sounds like there may be a hybrid MBR which isn't synced with the same values from the GPT | 13:23 |
mgedmin | also, this is a VM, right? | 13:23 |
Mudchains | mgedmin yes | 13:24 |
TJ- | Mudchains: also... 'sda' means the first (SCSI) storage device discovered on boot, which *may not* be the same on each boot and may be different between host+chroot and pure host boots | 13:24 |
Mudchains | mgedmin generation 1 Hyper originaly, converted to vmware :) | 13:24 |
TJ- | Mudchains: what does "sgdisk -p -O /dev/sda" report ? | 13:27 |
Mudchains | TJ- MBR info ( i am in a console session @ vmware, so I cant copy stuff) | 13:28 |
Mudchains | TJ- 1 primairy partition, mbr, code 0XEE :) | 13:28 |
TJ- | Mudchains: you need to compare the GPT info with the MBR info shown by that command | 13:28 |
TJ- | Mudchains: if it covers the entire 'disk' that is a protective MBR, not hybrid | 13:29 |
mgedmin | wait, there are two disks here: the real physical one that vmware deals with, and the virtual disk used by the VM, which is probably mapped to a partition or a file by vmware | 13:29 |
mgedmin | what does 'vmware console session' mean? | 13:30 |
TJ- | I thought this was all being done inside the VM's view of the hardware | 13:30 |
mgedmin | emulated /dev/tty1 of the VM? or some CLI to the vmware host side? | 13:30 |
mgedmin | just making sure | 13:30 |
Mudchains | I am not in a ssh session or something. I am on a console session. | 13:30 |
Mudchains | like you are really behind the server ;) in a physical world | 13:31 |
Mudchains | mgedmin the real disks you can ignore, as its stored on a ISCI volume. I have 3 disks in the VM :) disk layout: /dev/sda = boot disk, /dev/sdb = / , /dev/sdc = swap | 13:33 |
mgedmin | do these disks have partition tables? | 13:34 |
mgedmin | (or are you ext4-formatting the entire block device?) | 13:34 |
mgedmin | (if there's no partition table there might not be space for grub itself) | 13:36 |
Mudchains | mgedmin yes the disks have each one, 1 partition. (mostly ext4 exept the swap disk) | 13:36 |
Mudchains | the /dev/sda (boot) disk has 1 partition, that is flagged bios_grub and is using all the space | 13:37 |
Mudchains | TJ- when I use gdisk -l its says Found valid GPT with protective MBR; using GPT | 13:43 |
TJ- | Mudchains: there's your problem. One partition flagged as bios_grub means the BIOS boot partition, which grub-install should write its core image into (this is NOT a file-system and is NOT /boot/ or /boot/grub/) | 13:53 |
TJ- | Mudchains: if you've not working in a chroot then it's likely grub-install's defaults for other options may be chosen incorrectly | 13:54 |
Mudchains | TJ- the sda1 is the one with the bios_grub flag :) | 13:54 |
TJ- | Mudchains: OK, so what is mounted at /mnt/boot/ ? | 13:55 |
TJ- | Mudchains: or is it the root-fs mounted at /mnt/ ? | 13:55 |
Mudchains | TJ- /dev/sda1 | 13:55 |
=== gry_ is now known as gry | ||
Mudchains | and /dev/sdb1 ( / in production vm) is mounted as /mnt | 13:56 |
TJ- | Mudchains: OK, that IS the problem then, as I said, bios_grub partition is NOT a file system and is NEVER mounted, it is a raw bitmap of core image | 13:56 |
TJ- | Mudchains: so what've you have happening is GRUB's bootstrap code in sector 0 (440 bytes) loading the bytes from sda1 into memory and handing execution to them, expecting it to be the binary core image ... but it isn't | 13:57 |
mgedmin | is it required? usually grub stores the core image in the empty space between the MBR and the first partition? | 13:57 |
TJ- | mgedmin: its GPT | 13:57 |
* mgedmin is not sure what happens with GPT and non-EFI boots | 13:57 | |
TJ- | mgedmin: there is no spare space | 13:57 |
mgedmin | ah! | 13:57 |
TJ- | mgedmin: GPT primary starts at sector 1 .. this is why the BIOS Boot partition was needed for GPT in BIOS boot mode | 13:58 |
Mudchains | TJ- do I need to resize the /boot partition? | 13:58 |
TJ- | no longer are there spare/unused sectors at start of disk that can be relied on | 13:58 |
mgedmin | TIL! | 13:58 |
mgedmin | can vmware do EFI boots? | 13:58 |
Mudchains | yes | 13:58 |
Mudchains | HyperV vm thats I converted today (generation 2) are booting using EFI | 13:59 |
TJ- | Mudchains: lets go back to basics since I'm confused. Is there a separate root-fs AND a separate /boot/ file-system in different file systems? | 13:59 |
TJ- | Mudchains: note sda1 bios_grub is NOT for a /boot/ file-system | 13:59 |
mgedmin | EFI boot need an ext4 /boot _and_ a vfat /boot/efi, so you're going to have to repartition /dev/sda in any case | 14:00 |
mgedmin | (a) MBR + /boot partition, (b) GPT + bios_grub + /boot partitions, (c) GPT + /boot + /boot/efi partitions | 14:00 |
mgedmin | are the options | 14:00 |
Mudchains | mgedmin I am not getting EFI options at boot, so VMware knows this VM didnt support EFI. At other VM's i get EFI options | 14:01 |
TJ- | rewrite (a) as "MBR + spare sectors from 1 to 2047 + /boot/ partition | 14:01 |
mgedmin | .. hey in theory can't the ext4 superblock be shifted forward enough to make space for the grub core image? | 14:01 |
TJ- | Mudchains: is there any reason that /boot/ is not in the root-fs ? | 14:02 |
mgedmin | no, I was misremembering something I read about in https://amos.me/blog/2019/reading-files-the-hard-way-3/ | 14:02 |
Mudchains | TJ- I have a disk for the boot partition (/dev/sda) a disk for the OS ( /dev/sdb) and a disk for the swap file (/dev/sdc) | 14:02 |
mgedmin | the first 1024 bytes are reserved for boot sectors etc, not eough for grub | 14:02 |
Mudchains | this is for performance purposes | 14:02 |
TJ- | Mudchains: how large is sda ? | 14:02 |
Mudchains | 1 GB | 14:02 |
TJ- | Mudchains: OK, so simply repartition sda, 2 partitions, 1 bios_grub (2MB) and the rest for for /boot/ ext4 filesystem ... sorted | 14:03 |
TJ- | Yes, that is 2 Megabyte! | 14:03 |
rapidwave | Is there a GUI tool for creating/configuring a theme for Ubuntu? | 14:03 |
rapidwave | Also, how do I check which DE I'm using? | 14:04 |
Mudchains | TJ- I must use the same command as before? | 14:04 |
Mudchains | from outside chroot? or withing chroot? | 14:04 |
mgedmin | rapidwave: echo $XDG_CURRENT_DESKTOP, or show us a screenshot and we'll make a guess | 14:04 |
TJ- | Mudchains: if we assume the /boot/ becomes sda2, then you'd "mount /devsda2 /mnt/boot" then "grub-install /dev/sda --boot-directory=/mnt/boot/" and it should work. | 14:05 |
Mudchains | TJ- must I mount /dev/sdb1 also as /mnt? Like before | 14:06 |
TJ- | Mudchains: what I think was happening is: grub-install was over-writing sda1 because it was flagged as bios_boot but was also mounted at /mnt/boot -- although I cannot see how it would not have failed since grub-install first copies the files into /boot/grub/ then writes the core image to bios_boot | 14:06 |
rapidwave | Okay. So, I'm using LXQt, that must be a version of Xfce | 14:07 |
TJ- | Mudchains: yes, root-fs needs mounting to /mnt/ then the /boot/ file-system to /mnt/boot/ | 14:07 |
Mudchains | ok :) This is the disk layout now: https://i.imgur.com/ljmNmLI.png | 14:07 |
TJ- | Mudchains: I think you might have more success this time | 14:08 |
Mudchains | TJ- commands are finished, rebooting now | 14:09 |
Mudchains | TJ- I am getting a grub interface, waiting for input -.-' | 14:10 |
Mudchains | TJ- ah shit, i didnt fixed my /etc/fstab ... :) | 14:10 |
Mudchains | the /boot has a new UUID now | 14:11 |
TJ- | Mudchains: always something :) | 14:17 |
Mudchains | TJ- the fstab has the correct UUID now, but still grub rescue screen. Will try boot-repair now | 14:17 |
mgedmin | does one need to update-initramfs after updating the /boot mountpoint in /etc/fstab? | 14:18 |
=== Etua_ is now known as Etua | ||
Mudchains | boot-repair is complaining about the boot_grub & gpt again.. :) hehe | 14:19 |
mgedmin | when you say "grub rescue screen", what do you mean? | 14:20 |
Mudchains | mgedmin it says GRUB GNU... and there is a console screen of grub | 14:21 |
Mudchains | where I can type cmds in :) | 14:21 |
mgedmin | but no boot menu? | 14:21 |
Mudchains | indeed | 14:21 |
mgedmin | are there any errors? | 14:21 |
Mudchains | no | 14:21 |
mgedmin | it's as if it doesn't know where to look for the grub.cfg | 14:21 |
Mudchains | mgedmin this is the disk layout now @ live session: https://i.imgur.com/BTl6YGB.png | 14:22 |
mgedmin | can you ls (hd0,1) and see the contents of your /boot? | 14:22 |
mgedmin | or was it (hd0,2)? I don't remember if partition numbers are 0-based or 1-based in grub | 14:22 |
Mudchains | mgedmin can I do that @ that grub cli ? | 14:22 |
mgedmin | yes, the grub cli is very versatile | 14:23 |
TJ- | Mudchains: grub rescue> is the core image loaded but is not finding the /boot/ so use "set" and check what "root" and "prefix" are set to | 14:23 |
mgedmin | you can also use ls to look inside subdirectories, and then load your actual grub.cfg with the configfile command | 14:24 |
TJ- | Mudchains: use "ls" to list devices and partitions and e.g. "ls (hd0,2)/" to hopefully see the content of sda2 (if sda == hd0) | 14:24 |
mgedmin | which should give you the grub menu from your ubuntu install | 14:24 |
mgedmin | where you should be able to boot and repair your grub from the actual system (rather than a live session) | 14:24 |
mgedmin | which might work better maybe hopefully? | 14:24 |
Mudchains | https://i.imgur.com/381QtIB.png | 14:24 |
TJ- | Easiest is to just use "set prefix=(hd0,2)/grub" if hd0,2 is /boot/, and then do "insmod normal" then "normal" | 14:25 |
multifractal | Does anybody know how you can get a dock on a second display in 18.04? There is only a dock on the primary display for me. | 14:25 |
TJ- | Mudchains: check what "ls (hd0,gpt2)/" reports | 14:25 |
Mudchains | lost+found and grub | 14:25 |
mgedmin | also listen to TJ-, don't listen to me, I forget people don't use MBR | 14:25 |
TJ- | Mudchains: Good... then do "ls (hd0,gpt2/grub/" - do you see directories and grub.cfg ? | 14:26 |
TJ- | oops, typo | 14:26 |
TJ- | Mudchains: Good... then do "ls (hd0,gpt2)/grub/" - do you see directories and grub.cfg ? | 14:26 |
TJ- | My bet is there's no grub.cfg | 14:26 |
Mudchains | there is no grub.cfg | 14:26 |
TJ- | aha, so let's do it manually (or try) | 14:27 |
Mudchains | i386-pc, local , fonts, grubev folders/files | 14:27 |
TJ- | Mudchains: "ls (hd0,gpt2)/" and show us the listing in a screenshot | 14:27 |
* TJ- wonders if the kernel images/initrd are missing | 14:27 | |
Mudchains | TJ- https://i.imgur.com/169SEtk.png | 14:28 |
TJ- | Mudchains: there's your problem, no kernels installeed | 14:29 |
TJ- | Mudchains: my bet is they're in the root-fs instead | 14:29 |
mgedmin | there was earlier talk about deleting the /boot partition? but then Mudchains said he was going to copy it from some backup so it wouldn't be empty again? | 14:29 |
mgedmin | yeah, maybe ls (hd1,gpt1)/boot and see if it's not empty? | 14:30 |
TJ- | Mudchains: try "ls (hd1,gpt1)/boot/" -- do you see vmlinuz* and initrd.img* files? | 14:30 |
TJ- | Mudchains: if not try "ls (hd2,gpt1)/boot/" for the same | 14:30 |
mendi | hi does the nouveau driver work with wine (which requires dxvk)? | 14:30 |
TJ- | Mudchains: if you see some files please screenshot it for me | 14:30 |
Mudchains | TJ- https://i.imgur.com/rYNfwfn.png | 14:32 |
TJ- | Mudchains: OK so we can try a manual boot from hd1,gpt1 | 14:32 |
TJ- | Mudchains: "insmod linux" | 14:33 |
Mudchains | TJ- that gives a new cli line | 14:33 |
TJ- | Mudchains: then "linux (hd1,gpt1)/boot/vmlinuz root=/dev/sdb1 ro" then, once it has loaded into memory, do "initrd (hd1,gpt1)/boot/initrd.img" then finally, once that is in memory, "boot" | 14:34 |
TJ- | I think I have that syntax correct but not needed to do it in a while | 14:35 |
TJ- | Mudchains: it may be it starts Linux but then gets stuck in the initialramfs shell, from where we can further 'help' it along, hopefully | 14:35 |
Mudchains | file /boot/vmlinuz not found | 14:35 |
TJ- | Mudchains: hmmm, I assumed the entries in the "ls (hd1,gpt1)/" listing were symlinks to files in /boot/ ... maybe they're real files ? | 14:36 |
Mudchains | TJ- ls (hd1,gpt1)/boot/ gives no files/directories back | 14:37 |
TJ- | Mudchains: try "linux (hd1,gpt1)/vmlinuz root=/dev/sdb1 ro" | 14:37 |
TJ- | Mudchains: this'll fail if vmlinuz is only a dangling symlink... in which case you didn't install kernel images or run update-initramfs | 14:38 |
Mudchains | TJ- : https://i.imgur.com/Pf0lNML.png | 14:38 |
TJ- | Mudchains: hmm, so no kernel installed! also, wouldn't have made a difference but should be root= not root/ | 14:39 |
TJ- | Mudchains: you'll have to fix it from a chroot on the host, can't get any further without a kernel and initrd.img | 14:40 |
Mudchains | ok :) I will boot into live cd again | 14:40 |
Mudchains | TJ- I have to follow https://help.ubuntu.com/community/Grub2/Installing#via_ChRoot ? | 14:43 |
Gusj | Hello everyone, I have installed ubuntustudio 19.04, lubuntu 18.04lts and tried ubuntu 16.04, and in all is the same problem, laptop trackpad and keyboard do not work, I think I have tracked it down with the logs to the following: | 14:43 |
Gusj | 1.482768] i8042: PNP: PS/2 Controller [PNP0303:KBC,PNP0f13:MOUE] at 0x60,0x64 irq 1,12 | 14:44 |
Gusj | [ 1.485667] serio: i8042 KBD port at 0x60,0x64 irq 1 | 14:44 |
Gusj | [ 1.485677] serio: i8042 AUX port at 0x60,0x64 irq 12 | 14:44 |
Gusj | [ 1.540486] input input1: hash matches | 14:44 |
TJ- | Mudchains: those instructions are not optimal | 14:44 |
Gusj | [ 1.831655] hidraw: raw HID events driver (C) Jiri Kosina | 14:44 |
Gusj | [ 2.817045] i8042: Can't write CTR while closing AUX port | 14:44 |
TJ- | Mudchains: first check the file-systems are on the same devices as earlier with "sudo blkid" | 14:44 |
TJ- | Mudchains: assuming sda2 is /boot/ and sdb1 is root-fs, then do: | 14:45 |
Gusj | Its at the kernel level, a beginner here, but I have been learning a lot through this bug.. have tried many many commands in the grub file, and many online solutions but have not been successful, the laptop keyboard and trackpad work perfectly in the machine bios, is there any way that i can FORCE the i8042 AUX PORT? | 14:46 |
TJ- | Mudchains: "sudo -i" then "mkdir /target; mount /dev/sdb1 /target; chroot /target mount -a" then do "mount | grep target" and confirm that there is sda2 on /target/boot/ | 14:46 |
TJ- | Mudchains: if so, then continue with "for n in proc sys dev etc/resolv.conf; do mount --rbind /$n /target/$n; done" then "chroot /target apt install linux-image-generic" | 14:47 |
Mudchains | TJ- https://i.imgur.com/qJfGzUD.png | 14:47 |
TJ- | Mudchains: this should install the kernel and cause the initrd.img to be built, after which do "chroot /target update-grub" and that should generate the /target/boot/grub/grub.cfg file | 14:47 |
TJ- | Mudchains: ahhh... so there's no correct entry in /target/etc/fstab for the /boot/ device. Can you do "cat /target/etc/fstab; blkid" and screenshot it ? | 14:48 |
Mudchains | TJ- in chroot or outside? | 14:49 |
TJ- | Mudchains: outside, just as I typed the commands. The fact I include /target/ infers the command is from outside the chroot | 14:49 |
TJ- | Mudchains: since the broken install is mounted under /target/ | 14:50 |
TJ- | Mudchains: actually, did you for the "for n in ..." command ? | 14:50 |
Mudchains | TJ- not yet | 14:51 |
TJ- | Mudchains: I'd have expected the 'grep' to report /target/proc /target/sys /target/dev and so on | 14:51 |
TJ- | Mudchains: aha! that's the cause! | 14:51 |
TJ- | Mudchains: I was premature in saying to do ".... mount -a" ! | 14:51 |
TJ- | Mudchains: so, do the "for n in ..." THEN do "chroot /target mount -a" and then "mount | grep target" | 14:52 |
TJ- | Mudchains: "mount -a" causes all entries in fstab to be auto-mounted | 14:52 |
beterraba | hello all. i'm trying to install something without `sudo apt-get blablabla`. this is the library that i need: https://github.com/Z3Prover/z3/releases | 14:52 |
beterraba | i find there's a `ubuntu` release. i tried to look in the website, but i simply don't know how to proceed | 14:53 |
Gusj | Where I think the problem lies is in the following lines, sorry for the paste before: i8042: Can't write CTR while closing AUX port, i8042: Can't reactivate AUX port | 14:54 |
Mudchains | TJ- now doing the for loop? https://i.imgur.com/0oxjZ6y.png | 14:54 |
Mudchains | I dont know if I am in chroot or outside now -.- | 14:55 |
TJ- | Mudchains: yes, do the "for..." command now ... type commands exactly as I gave them, from the host. I'm not causing you to enter the chroot permanently, only execute commands in it one at a time from outside | 14:56 |
Mudchains | thanks | 14:56 |
TJ- | Mudchains: when you do "chroot /target <some command>" it only runs <some command> inside the chroot but 'you' stay outside | 14:57 |
TJ- | Mudchains: whereas "chroot /target" would but 'you' inside it until you typed 'exit' | 14:57 |
lordcirth | Though you can of course use 'bash' as the command to get a shell inside it as well. | 14:57 |
TJ- | s/but/put/ | 14:57 |
Mudchains | TJ- i am getting: https://i.imgur.com/Alc6JX9.png | 14:58 |
TJ- | beterraba: see https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html and the various i8042.* options | 14:58 |
beterraba | thank you ! | 14:59 |
beterraba | ket me take a look | 14:59 |
TJ- | Mudchains: "for n in proc sys dev etc/resolv.conf; do sudo mount --rbind /$n /target/$n; done" | 14:59 |
Gusj | Is there a way to force the kernel to activate the i8042 AUX port? | 14:59 |
TJ- | beterraba: Ouch, sorry, I gave you a link meant for Gusj!! | 14:59 |
TJ- | Gusj: see https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html and the various i8042.* options | 14:59 |
beterraba | ah ok | 15:00 |
Gusj | @TJ- ahh thank you :) let me check that out now.. | 15:00 |
Mudchains | TJ- next step? https://i.imgur.com/BDTgHo2.png | 15:01 |
TJ- | Gusj: I don't think there's a 'direct' way but I seem to recall that i8042.noloop" and/or .nopnp might have the same effect | 15:01 |
ghost2911 | hi. how to list installed packages so that it lists not in column but line but horizontally like paragraph each package separated by space | 15:02 |
Gusj | @TJ- thank you will try some of them right now.. | 15:02 |
TJ- | Mudchains: OK, target to knowhere means the DNS resolver on the live host is different to whats in the broken system.. it may cause problems with the chroot not being able to resolve hostnames and therefore not be able to download packages. We'll soon find out | 15:02 |
TJ- | Mudchains: so "chroot /target mount -a" and then check for /target/boot/ with "mount | grep target" | 15:03 |
Mudchains | TJ- yep, the live host has different ipadres etc, because else apt-get isnt working the live-cd session | 15:03 |
Gusj | @TJ- It is very strange they both work perfectly in the bios, and the keyboard works in the grub menu as well, but when you get to the login screen, both are gone.. | 15:04 |
TJ- | Gusj: hang on... is it a PS/2 keyboard or USB? | 15:04 |
Mudchains | TJ- the mount | grep target gives a lot of folders | 15:05 |
TJ- | Mudchains: good :) | 15:05 |
TJ- | Mudchains: do you see /target/boot mentioned? | 15:05 |
mgedmin | Mudchains: does it list /target/etc/resolv.conf ? | 15:05 |
TJ- | /target/etc/resolv.conf will be there but is a dangling sym-link so we may have to remove it to unmask what it mounted over | 15:05 |
mgedmin | you need /target, /target/boot, /target/dev, /target/proc, /target/sys and /target/etc/resolv.conf | 15:06 |
Gusj | @TJ- it is the laptops internal keyboard and trackpad, I think yes PS/2 this is how the kernel sees them before giving the AUX port error, 1.482768] i8042: PNP: PS/2 Controller [PNP0303:KBC,PNP0f13:MOUE] at 0x60,0x64 irq 1,12 | 15:06 |
Mudchains | sda2 is mounted as /target/boot | 15:06 |
TJ- | Gusj: Yes, I concur, PS/2 | 15:06 |
TJ- | Mudchains: Yay! | 15:06 |
Mudchains | there is no resolv.conf | 15:06 |
TJ- | Mudchains: right, lets undo the resolv.conf issue with "umount /target/etc/resolv.conf" | 15:07 |
TJ- | Mudchains: if that says it doesn't exist that is fine | 15:07 |
Gusj | @TJ- I have been messing around trying to resolve this with external USB kybrd and mouse, USB works perfectly. | 15:07 |
Mudchains | TJ- no mount point speci... | 15:07 |
TJ- | Gusj: are you able to pastebin the entire kernel log, e.g. "pastebinit <( journalctl -k )" using the USb keyboard, or SSH session? | 15:08 |
TJ- | Mudchains: that's fine, ignore for now, lets test if the chroot has DNS capability | 15:08 |
TJ- | Mudchains: "chroot /target ping -c 5 1.1.1.1" | 15:08 |
TJ- | Mudchains: if that works try "chroot /target ping iam.tj" | 15:09 |
TJ- | Mudchains: hopefully the name lookup will work. if not, we need to fix that | 15:09 |
Mudchains | ip works, dns not | 15:09 |
Gusj | @TJ- yes already have it up since last night, here it is the Xorg.0.log: https://paste.ubuntu.com/p/fQZYZyDyR4/ thank you for taking the time to help me | 15:09 |
TJ- | Mudchains: thought so. What does "grep nameserver /etc/resolv.conf" report ? | 15:10 |
Mudchains | 127.0.0.53 | 15:10 |
Mudchains | isnt it a netplan setting? | 15:10 |
TJ- | Mudchains: that points to the local systemd-resolved so now check what is in the chroot, with "cat /target/etc/resolv.conf" (which may not exist!) | 15:11 |
TJ- | Gusj: that's the Xorg log; it'd be more useful to see the kernel log as I indicated above | 15:12 |
Mudchains | TJ- no such file | 15:12 |
TJ- | Mudchains: OK, let's try to manually bodge this but you'll need to fix it up later once the system boots | 15:12 |
Gusj | @TJ- Ahh I am sorry for the confusion, where is the log that you specify located? | 15:12 |
mgedmin | augh https://unix.stackexchange.com/questions/429724/how-do-i-mount-a-file-on-top-of-a-broken-symbolic-link as no answer | 15:12 |
mgedmin | /as/has | 15:12 |
TJ- | Gusj: just issue the command "pastebinit <( journalctl -k )" and give us the URL | 15:12 |
TJ- | Mudchains: "echo 'nameserver 127.0.0.53' > /target/etc/resolv.conf" then retry "chroot /target ping iam.tj" | 15:13 |
dbtid | i'm running hostapd and dnsmasq to provide a loal AP at home. i have a dozen or so small machines that use this AP. how do I get dnsmasq to bind the hostnames of the WiFi client machines so i don't have to have large /etc/hosts files on all the machines? dnsmasq knows the hostnames (they're in /var/lib/misc/dnsmasq.leases). | 15:13 |
Gusj | @TJ- here it is: http://paste.ubuntu.com/p/5y2bBgCyqX/ | 15:14 |
Mudchains | TJ- https://i.imgur.com/RNCCz3p.png | 15:15 |
mgedmin | TJ-: /target/etc/resolv.conf is most likely a symlink to ../run/systemd/resolve/stub-resolv.conf; /target/run/ is empty so the echo will likely fail with ENOFILE | 15:15 |
mgedmin | it's _annoying_ that bind mounts don't work when the target is a broken symlink | 15:16 |
TJ- | mgedmin: yes, that's what I would expect | 15:16 |
mgedmin | mkdir /target/run/systemd/resolve seems like overkill, just to touch a stub-resolv.conf and bind mount it there? | 15:16 |
mgedmin | mount --bind /run /target/run seems dangerous maybe? | 15:16 |
TJ- | Mudchains: that suggests /etc/ doesn't exist! does "ls /target/etc/" list files? | 15:16 |
mgedmin | mv /taget/etc/resolv.cfg && cat /etc/resolv.cfg > /target/etc/resolv.cfg seems unclean and dangerous if not undone before the reboot | 15:17 |
TJ- | mgedmin: I tend to avoid it since it mounts info about HOST services into the chroot so if chroot starts its own services (due to package upgrades/installs) it causes issues for the host | 15:17 |
TJ- | mgedmin: also prevents /target/run/ from being unmounted | 15:17 |
TJ- | The echo redirect should work since > creates a file, unless /target/etc/ directory is missing | 15:18 |
mgedmin | TJ-: why do you say /etc/ doesn't exist? /target/etc/resolv.conf is a broken symlink, echo > will try to follow the symlink into a non-existent directory and try to create a file there, which fails because the parent directory is missing | 15:18 |
Mudchains | arg | 15:18 |
mgedmin | the parent directory being /target/run/systemd/resolve/ | 15:18 |
Mudchains | the console session crashed.. I cant do anything atm | 15:18 |
TJ- | mgedmin: oh I see what you're getting out. My thought process was: the mount --rbind reported a dangling symlink due to it creating the rbind... but I see what you mean now, it tried to follow an existing /target/etc/resolv.conf to possibly /run/ which isn't there | 15:19 |
TJ- | Mudchains: what does "readlink -e /target/etc/resolv.conf" report ? | 15:20 |
mgedmin | yeah, google tells me you can't bind-mount anything over a symlink, the bind mount always resolves the link and mounts on top of its target (which fails if the target doesn't exist) | 15:20 |
mgedmin | https://serverfault.com/questions/322906/how-do-i-do-a-bind-mount-over-a-symlink | 15:20 |
TJ- | mgedmin: sorry; trying to juggle too many things here | 15:20 |
mgedmin | np, just trying to help! | 15:20 |
mgedmin | you know more than I do | 15:20 |
TJ- | yeah, but I can be blind at times; need keeping on track! | 15:21 |
TJ- | the more the merrier as Robin Hood used to say :) | 15:21 |
mgedmin | I could probably fix this myself, but I'm struggling to explain how in a clear and unambiguous way | 15:21 |
TJ- | Mudchains: let us know when you've regained controlled | 15:21 |
TJ- | mgedmin: it's a pain when all we want to do is "chroot /target apt install linux-image-generic && chroot /target update-grub" ! | 15:22 |
TJ- | although I think update-grub isn't needed, should be triggered by the kernel package install | 15:22 |
Mudchains | TJ- control is back | 15:23 |
TJ- | problem is if those packages are installed but their files are missing, we'd need to force a --reinstall then | 15:23 |
TJ- | Mudchains: OK ... "readlink -r /target/etc/resolv.conf" | 15:23 |
TJ- | grrr, "-e" not "-r" !! | 15:23 |
mgedmin | yup! hey maybe /var/cache/apt/archives still has a cached kernel deb so dns wouldn't be needed? | 15:23 |
TJ- | mgedmin: hope so! | 15:23 |
Mudchains | TJ- its blanc | 15:23 |
mgedmin | "all components must exist" says the man page about readlink -e, so blank output is not surprising | 15:24 |
* mgedmin would go with 'mv /target/etc/resolv.conf /target/etc/resolv.conf.orig; echo nameserver 127.0.0.53 > /target/etc/resolv.conf' and then undo later right before/after rebooting into the actual system | 15:25 | |
TJ- | Mudchains: OK, ^^^^ | 15:26 |
TJ- | mgedmin: is saving me typing :) | 15:26 |
TJ- | Mudchains: once you've done the echo, try once more "chroot /target ping iam.tj" and lets hope it works! | 15:27 |
Mudchains | yep thats working! | 15:27 |
mgedmin | now to figure the version of the kernel package to install | 15:28 |
mgedmin | chroot /target dpkg -l 'linux-image-*' | 15:28 |
mgedmin | chroot /target dpkg -l 'linux-image-*' | grep ^i | 15:28 |
Mudchains | 4.15.0-64-generic (uname -r) | 15:28 |
mgedmin | uname -r will tell you the kernel in your live system | 15:29 |
mgedmin | not the one that was installed in /target before the /boot went away | 15:29 |
Mudchains | this is from production system | 15:29 |
mgedmin | ah right you have a copy still working, I keep forgetting | 15:29 |
mgedmin | chroot /target apt install --reinstall liux-image-4.15.0-64-generic | 15:30 |
mgedmin | only spell linux-image correctly ;) | 15:30 |
Mudchains | mgedmin https://i.imgur.com/CNwmaJ8.png | 15:31 |
mgedmin | okay, this means you had grub 1 installed in the system, but now with the live session you've set up grub 2 | 15:32 |
Gusj | @TJ- the lines that start repeating after line 806, I also see in the ttys when I switch to them with ctrl+alt+fX, when i try to write or no in any terminal, that output from kernel gets printed continually, 1 ACPI Bios error, and 4 ACPI errors | 15:33 |
mgedmin | Mudchains: I would do a chroot /target apt install grub-pc, unless TJ- has a better idea | 15:33 |
Mudchains | hmm ok :) | 15:34 |
TJ- | sorry, was away investigating a leak. Yes, I concur | 15:36 |
TJ- | that grub-legacy-ec2 stuff is so misleading | 15:36 |
Mudchains | so a chroot /target apt install grub-pc it is :) ? | 15:37 |
TJ- | Yes | 15:37 |
Guest_8 | hello, I need some help with recovering encrypted home dir | 15:37 |
TJ- | Guest_8: "ecryptfs-recover-private /home/.ecryptfs/$USER/.Private" you mean? | 15:37 |
Guest_8 | yes I think so i am on the same machine logged in to a livecd | 15:38 |
Mudchains | TJ- https://i.imgur.com/4HxPKbL.png | 15:38 |
Guest_8 | ecryptfs-utils doesnt seem to be installed on the live cd | 15:38 |
mgedmin | Guest_8: you can sudo apt install stuff in the live session (as long as they fit in RAM -- there's a writeable ramdisk) | 15:39 |
Guest_8 | yeah i tried that but doesnt find the file to install | 15:39 |
Guest_8 | ecryptfs-utils has no release candidate | 15:40 |
mgedmin | ah, you have to sudo apt update first, forgot about it | 15:40 |
TJ- | Mudchains: "chroot /target apt remove grub-legacy-ec2" | 15:40 |
Guest_8 | still doesnt install after update i tried that already | 15:41 |
mgedmin | (the livecd doesn't have /var/lib/apt/lists, and if it did, they would become outdated soon anyway) | 15:41 |
Guest_8 | how do i update that? | 15:41 |
TJ- | Guest_8: you may also need to do, prior to ecryptfs-recover-private, "sudo keyctl link @u @s" | 15:41 |
mgedmin | huh, is ecryptfs-utils not in main? the live session probably doesn't have universe in /etc/apt/sources.list | 15:41 |
TJ- | Guest_8: which release are you using? | 15:41 |
Guest_8 | 19.04 | 15:41 |
TJ- | !info ecryptfs-utils | 15:41 |
ubottu | ecryptfs-utils (source: ecryptfs-utils): ecryptfs cryptographic filesystem (utilities). In component universe, is optional. Version 111-0ubuntu5 (bionic), package size 103 kB, installed size 548 kB | 15:41 |
Mudchains | TJ- https://i.imgur.com/WZICitO.png | 15:42 |
TJ- | Mudchains: looking good! now do "chroot /target apt -f install" to clean up any package issues | 15:42 |
mgedmin | Mudchains: looks like everything's fine now! to double check ls /target/boot -- you should see a kernel; and ls /target/boot/grub -- you should see a grub.cfg | 15:42 |
TJ- | Mudchains: then finally do "chroot /target update-grub" | 15:43 |
mgedmin | apt -f install can't hurt | 15:43 |
TJ- | I think the -f will trigger update-grub and generate grub.cfg but best to be sure | 15:43 |
dbtid | i'm running hostapd and dnsmasq to provide a loal AP at home. i have a dozen or so small machines that use this AP. how do I get dnsmasq to bind the hostnames of the WiFi client machines so i don't have to have large /etc/hosts files on all the machines? dnsmasq knows the hostnames (they're in /var/lib/misc/dnsmasq.leases). | 15:43 |
Mudchains | TJ- mgedmin https://i.imgur.com/7qe5TSl.png | 15:43 |
mgedmin | I'd have expected the apt remove grub-legacy-ec2 to have done the dpkg --configure --pending at the end - it was processing kernel triggers after all | 15:44 |
TJ- | Mudchains: "chroot /target update-grub" | 15:44 |
mgedmin | so, kernel - yes, grub.cfg -- no; try the update-grub that TJ- suggested | 15:44 |
Mudchains | mgedmin TJ- succesfully done :) | 15:45 |
TJ- | good to know update-grub won't create a menuentry for a vmlinuz that doesn't have a matching initrd.img | 15:45 |
Mudchains | found the image that we installed by apt install | 15:45 |
TJ- | Mudchains: good... now undo the /target/etc/resolv.conf change from earlier | 15:45 |
TJ- | Mudchains: "mv /target/etc/resolv.conf{.orig,}" | 15:45 |
TJ- | I hope I recall the backup file's name correctly there | 15:46 |
mgedmin | yes .orig | 15:46 |
Guest_8 | how can i add universe to the repos on a livecd? | 15:47 |
TJ- | Guest_8: "sudo add-apt-repository universe" | 15:47 |
mgedmin | oh nice! | 15:47 |
Guest_8 | thanks tj | 15:47 |
TJ- | Mudchains: in theory you've good to reboot now | 15:47 |
Mudchains | TJ- delete the resolv.conf and mv the .orig file? | 15:48 |
TJ- | Mudchains: "mv /target/etc/resolv.conf{.orig,}" | 15:48 |
mgedmin | no need to delete, mv will overwrite it | 15:48 |
TJ- | Mudchains: that ^^^ does it for you | 15:48 |
Mudchains | ok | 15:48 |
TJ- | Mudchains: the {.orig,} expands the command to read "mv /target/etc/resolv.conf.orig /target/etc/resolv.conf" | 15:49 |
mgedmin | I would do a sync; sync; sync; for i in dev proc sys boot; do umount /target/$i; done, but I'm paranoid | 15:49 |
TJ- | mgedmin: systemd/kernel will handle all that on shutdown :) | 15:49 |
Mudchains | TJ- didnt know that! thanks | 15:49 |
Mudchains | ok the mv cmd has done | 15:49 |
Mudchains | now reboot? | 15:49 |
mgedmin | (once I booted an ancient RH rescue floppy, recover some files on my hard disk, and did a 'reboot', only to discover that wasn't unmounting filesystems cleanly in that ancient rescue system) | 15:49 |
TJ- | Mudchains: yes "systemctl reboot" | 15:49 |
TJ- | Mudchains: and of course remove the live ISO | 15:50 |
TJ- | mgedmin: ahhh, the gold ole days :) | 15:50 |
TJ- | gold? good?! | 15:50 |
mgedmin | it was a valuable lesson | 15:50 |
TJ- | yes, like dd's conv=fdatasync,fsync before pulling a USB :) | 15:51 |
mgedmin | remember when IDE hard disks were called /dev/hda and USB flash disks were called /dev/sda? and then it changed one day? and I did dd if=/dev/zero of=/dev/sda without checking? | 15:51 |
Mudchains | its booting! | 15:51 |
Mudchains | :D :D | 15:51 |
mgedmin | \o/ | 15:51 |
sarnold | mgedmin: fun fun | 15:51 |
Mudchains | thank you so much TJ- and mgedmin \o/ | 15:52 |
Mudchains | if I had gold, I would gave it to you :D | 15:52 |
mgedmin | luckily I had an identical server next to it so I could mkfs and restore the destroyed root partition (it was sda1 not sda now that I rememebr) with tar and ssh | 15:52 |
Mudchains | Have a nice evening TJ- mgedmin | 15:59 |
TJ- | mgedmin: ahh yes, the switch inside libata hd>sd when it adopted the SCSI layey | 16:00 |
Gusj | @TJ- Hola.. TJ, where you able to take a look at the log file? | 16:43 |
TJ- | Gusj: the root cause is "atkbd serio1: keyboard reset failed on isa0060/serio1" as you surmised; it isn't clear why it is failing | 16:46 |
TJ- | Gusj: presumably it fails like this without any atkbd. or i8042. options on cmdline ? | 16:47 |
Gusj | @TJ- Thank you, I will delete all the commands in the grub and test it only with 'quiet splash' right? | 16:48 |
TJ- | Gusj: I'd lose the quiet splash too so you can see what is going on :) I hate having a GUI getting between me and the real system! | 16:48 |
Gusj | @TJ- Got it thank you will do that right now, and report back | 16:49 |
Gusj | @TJ- Just tried it with no commands and when I try to search for serio1 or atkbd I find nothing in the new log: http://paste.ubuntu.com/p/xhbnMKWVtK/ | 16:53 |
Gusj | @TJ- The only thing that I have noticed that does something different, is when i used the atkbd reset comm in Grub, the keyboard did not work but the numlock light would be off, anyother way, including now, the numlock light stays permanently on | 16:55 |
Gusj | @TJ- If the laptop keyboard and trackpad work in the bios, and in the Grub Menu, the problem can't be hardware right? or can it still be hardware related? | 16:56 |
TJ- | Gusj: line 539 shows it "i8042: PNP: PS/2 Controller [PNP0303:KBC,PNP0f13:MOUE] at 0x60,0x64 irq 1,12" so detection is working... config not so much! | 16:57 |
ioria | Gusj, already tried 'i8042.direct i8042.dumbkbd' parameters ? | 16:58 |
TJ- | Gusj: enable debug logging with "i8042.debug=1" so you get detail of the i8042 operations | 17:01 |
TJ- | Gusj: looks like this bug report covers it and there's a patch further down that appears to solve it, not sure where that had got to in mainline though https://bugzilla.kernel.org/show_bug.cgi?id=195471 | 17:04 |
ubottu | bugzilla.kernel.org bug 195471 in Input Devices "Keyboard not working after resume with kernel 4.10.5" [Normal,New] | 17:04 |
TJ- | Gusj: you're getting this on a cold boot though so it seems to affect that C5 more seriously | 17:08 |
TJ- | Gusj: since that report mentions suspend/resume that infers ACPI, so let's take a punt on an ACPI OSI workaround. Check out https://iam.tj/prototype/enhancements/Windows-acpi_osi.html | 17:10 |
=== Etua_ is now known as Etua | ||
=== jcara_ is now known as jcara | ||
B|ack0p | hi. i am getting this error everytime i install something or check apt update: libdvd-pkg: `apt-get check` failed, you may have broken packages. Aborting... | 18:01 |
B|ack0p | how can i fix it? | 18:01 |
lordcirth | B|ack0p, 'apt install -f' ? | 18:02 |
B|ack0p | lordcirth: let me check | 18:03 |
B|ack0p | lordcirth: 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. | 18:03 |
B|ack0p | nothing happened | 18:03 |
lordcirth | B|ack0p, ok. So it's only when attempting to change something that you have a problem. What Ubuntu version? Was it upgraded from a previous version? Any 3rd part repos? | 18:04 |
B|ack0p | 18.04 | 18:04 |
B|ack0p | no it is fresh install | 18:04 |
lordcirth | B|ack0p, can you pastebin the full output of 'apt update'? | 18:05 |
B|ack0p | ok | 18:05 |
B|ack0p | lordcirth: https://paste.ubuntu.com/p/3X9xF4tw67/ | 18:06 |
B|ack0p | it didnt give the error now | 18:06 |
B|ack0p | it happened while i was installing something | 18:06 |
lordcirth | B|ack0p, ok, what were you trying to install? | 18:07 |
B|ack0p | lordcirth: well i installed many things | 18:07 |
B|ack0p | gnome shell extensions | 18:07 |
B|ack0p | tweak tool | 18:07 |
B|ack0p | codecs | 18:07 |
lordcirth | B|ack0p, ok, and did they all fail? | 18:07 |
B|ack0p | lordcirth: now i tried to install "htop" and it gave error | 18:07 |
B|ack0p | let me pastebin | 18:07 |
B|ack0p | lordcirth: https://paste.ubuntu.com/p/BFg2YqCWnj/ | 18:08 |
B|ack0p | this is full log | 18:08 |
lordcirth | B|ack0p, first result of a search: https://askubuntu.com/questions/1094062/libdvd-pkg-apt-get-check-failed-you-may-have-broken-packages-aborting | 18:09 |
B|ack0p | lordcirth: i think it happened while i was installing this: https://websiteforstudents.com/how-to-install-video-audio-codecs-on-ubuntu-18-10-18-04-16-04-lts/ | 18:09 |
B|ack0p | thx i am trying again | 18:10 |
B|ack0p | lordcirth: installed without error.. | 18:12 |
B|ack0p | thx | 18:12 |
lordcirth | great. | 18:12 |
B|ack0p | could you please tell me a package necessary to install to test? :p | 18:13 |
B|ack0p | i wanna check again if i will get any error while apt get install ... | 18:13 |
lordcirth | B|ack0p, Seems fixed to me. That answer worked for other people. But you could install 'tmux'. | 18:15 |
B|ack0p | what does it do? | 18:16 |
Gusj | @TJ- Hi there, Sorry that I left for a bit.. thank you for the links you sent, looking at them right now to see if I understand what to do next.. | 18:16 |
B|ack0p | lordcirth: done without error; https://paste.ubuntu.com/p/cxcfJsqb4S/ | 18:17 |
B|ack0p | thx a lot! | 18:17 |
Gusj | @ioria Hi there, thank you have not tried only with those two (2) parameters you just sent, will do.. | 18:19 |
kinghat | you guys know of an SSD tester for speeds on linux? like crystal disk on windows. | 18:22 |
sarnold | fio and bonnie++ are packaged for ubuntu; I have trouble understanding their output usually (one is incomprehensible, the other difficult :) so I usually just use dd and either conv=fdatasync or or oflags=sync | 18:25 |
=== im0nde_ is now known as im0nde1 | ||
kinghat | i guess im not after the smart info im looking for performance measuring like: https://crystalmark.info/en/software/crystaldiskmark/ | 18:26 |
sarnold | fio and bonnie++ don't do smart; if you want smart data then you need to use the smartmontools package | 18:28 |
kinghat | smartmontools is a gooey? | 18:29 |
sarnold | I don't see any gui libraries in its dependencies | 18:29 |
lordcirth | It doesn't have it's own GUI, no. But many of the GUI disk utilities will read smart data. | 18:30 |
kinghat | ya im play with GSmartControl atm but also wondering if there is a gooey for benchmarking the drive like CrystalDiskMark for win. | 18:33 |
kinghat | wonder if GNOME Disks will bench ssds | 18:46 |
Gusj | @TJ- @ioria Just noticed that with 'i8042.direct i8042.dumbkbd' I can change the brightness of screen and activate/deactivate wifi with the special function keys FN+f2 f3 FN+f9 but that is it, no other function keys work, specially nos the one to activate deactivate trackpad, and no other key works in general, I think thaat I now understand why the video was assigned as a keyboard | 18:48 |
kinghat | pretty huge difference between samsung 830 and 840: https://irc.kinghat.info/uploads/d761afd525c20ea4/image.png https://irc.kinghat.info/uploads/4584a4b22cca3351/image.png | 19:18 |
=== BrianG61UK_ is now known as BrianG61UK | ||
kinghat | so yes, GNOME Disks does benchmarking 😛 | 19:18 |
sarnold | kinghat: I don't know what your workload is like, but 100MB is a *very* small sample size for these things | 19:26 |
sarnold | kinghat: quite often they'll have two gigs or four gigs of fast cache-ish space and then re-shuffle writes to slower storage later | 19:26 |
kinghat | i transferred 75GB file from the 840 to the 830 and it was transferring around 100ishMB/s | 19:27 |
kinghat | and that 830 write test pretty much confirms that. | 19:27 |
kinghat | was just trying to figure out what the transfer was so slow between two SSDs | 19:28 |
kinghat | looks like the culprit was the 830. | 19:28 |
Gusj | @TJ- @ioria Just noticed that with 'i8042.direct i8042.dumbkbd' I can change the brightness of screen and activate/deactivate wifi with the special function keys FN+f2 f3 FN+f9 but that is it, no other function keys work, specially nos the one to activate deactivate trackpad, and no other key works in general, I think thaat I now understand why the video was assigned as a keyboard | 19:29 |
TJ- | Gusj: are you sure the keyboard is disabled in consoles (TTYs) as well as in GUI? this could be an Xorg/libinput issue | 19:31 |
Gusj | @TJ- Yes it is disabled just checked now, it does not work there, in any tty I get the login and I can write with my usb keyboard, but the TTY's just keep outputting those lines that you can see in the log, 1 ACPI BIOS (bug) error, and 4 ACPI errors | 19:35 |
Gusj | trying the Acpi windows 20XX solutions working my way from 2013 to double check before trying to patch the kernel, | 19:35 |
Gusj | @TJ- I have not done this before, applying a patch to the kernel, can you recommend a how to that I can follow along to apply the patch in the link that you shared with me? | 19:36 |
TJ- | Gusj: good, because it is surprising how much that can fix, but, as that C55 looks to be quite old (2014?) I would think we'd know about any generic problems with that model | 19:36 |
TJ- | Gusj: first thing I'd do is test an older kernel as shown, so any earlier than 4.19 (4.15 from 18.04 for example) --- if that works we've got proif it is a regression. Then we can get the kernel team to do a git bisect and create kernels for you to test until we find the problem | 19:37 |
TJ- | Gusj: we also have mainline builds so you could test any mainline versions easily | 19:38 |
TJ- | !mainline | Gusj | 19:38 |
ubottu | Gusj: The kernel team supply continuous mainline kernel builds which can be useful for tracking down issues or testing recent changes in the Linux kernel. More information is available at https://wiki.ubuntu.com/Kernel/MainlineBuilds | 19:38 |
Gusj | @TJ- @ubottu I am sorry, but I am not understanding you on exactly I need to proceed, I know you say 4.15.. but I am in the dark.. | 19:43 |
TJ- | Gusj: the kernel bugzilla comments say that kernels older than 4.19 work... so the inference is that some regression was introduced at some point in or just after that release. So if you installed a mainline build of, say, 4.19, and it works, and then you install, say, 5.0, and it fails, we know some change there causes the regression | 19:45 |
ioria | Gusj, what's your release ? cat /etc/issue | 19:45 |
skyliner369 | How might I drop a marker in the journalctl -b logs? | 19:45 |
Gusj | @TJ- YEs that I understand, what I do not know is how to install the kernel you suggest.. | 19:45 |
TJ- | Gusj: at which point we can get the Ubuntu kernel team to build kernels between those 2 points (GOOD...BAD) to find the commit that causes it | 19:45 |
Gusj | @TJ- Thank you, here is my release Ubuntu 18.04.3 LTS \n \l | 19:46 |
TJ- | Gusj: oh... does the link not say? we have to manually download the kernel image .deb files and use dpkg -i X.deb to install them, then reboot and select | 19:46 |
ioria | Gusj, so you can install linux-generic that will bring 4.15 | 19:46 |
ioria | Gusj, sudo apt install linux-generic and from grub -> advanced oprion , choose 415 | 19:47 |
=== ben_r_ is now known as ben_r | ||
skyliner369 | I'm trying to track an issue that pops up after the desktop loads. | 19:48 |
Gusj | @TJ- @ioria jus tot make sure.. I am here: https://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15/ from here I will choose one from "Build for amd64 succeeded :" and take the 2nd one that says GENERIC "linux-headers-4.15.0-041500-generic_4.15.0-041500.201802011154_amd64.deb" I will donwload that, so far so good? | 19:52 |
ioria | Gusj, please, run apt cache policy linux-generic ; what you see ? | 19:53 |
ioria | Gusj, apt-cache policy linux-generic | 19:54 |
TJ- | Gusj: ^^^ | 19:54 |
Gusj | @ioria E: Invalid operation cache | 19:54 |
ioria | Gusj, apt-cache policy linux-generic | 19:54 |
Gusj | Installed: (none) | 19:54 |
Gusj | Candidate: 4.15.0.65.67 | 19:54 |
Gusj | Version table: | 19:54 |
Gusj | 4.15.0.65.67 500 | 19:54 |
Gusj | 500 http://do.archive.ubuntu.com/ubuntu bionic-updates/main amd64 Packages | 19:54 |
Gusj | 4.15.0.64.66 500 | 19:54 |
skyliner369 | for anyone who knows what they're looking at (shows up a ton in journalctl -b logs of sessions that show the desktop but lose USB and PS/2) Should I just nuke this install? https://usercontent.irccloud-cdn.com/file/86WPLttn/irccloudcapture3491313573036892024.jpg | 19:54 |
ioria | Gusj, right, so you can install 4.15 from the offcial repo with apt (as i stated above) | 19:55 |
ioria | Gusj, but you need to select it from the Grub screen | 19:56 |
Gusj | @ioria Thank you so I will do the following command? "sudo apt install linux-generic" in terminal, and then when it is finished reboot, then in Grub I need to choose advanced options and you said option 415? | 19:57 |
ioria | yes | 19:57 |
ioria | Gusj, 4.15 | 19:57 |
Gusj | @ioria Got it, and then the machine will boot up using that kernel, and my files will stay unaltered as this is only related to the kernel right? | 19:58 |
ioria | Gusj, yes | 19:58 |
Gusj | @ioria Thank you will do that | 19:59 |
ioria | Gusj, hope it works, gl | 19:59 |
Gusj | brb | 19:59 |
skyliner369 | How do I work the termbin pipe again? | 20:02 |
cocof | hi | 20:03 |
cocof | whats the best command to create new user with a home dir? | 20:03 |
cocof | useradd? | 20:04 |
Ben64 | adduser | 20:05 |
cocof | and to switch to it? su username? | 20:07 |
cocof | yes now it works :) | 20:07 |
cocof | strange useradd did not do much. | 20:07 |
skyliner369 | So with a log like this here https://termbin.com/5zfx should I just nuke this install? | 20:07 |
RonaldsMazitis | hello | 20:13 |
RonaldsMazitis | I have problem that my ubuntu does not show newly installed kernel | 20:14 |
RonaldsMazitis | grub menu has only old versions | 20:14 |
EriC^^ | RonaldsMazitis: type "ls -l /boot | nc termbin.com 9999" | 20:14 |
Lustic | Hi just installed 18.04, hangs at splash screen cant find the reason, anyone know ? | 20:14 |
EriC^^ | RonaldsMazitis: also "cat /boot/grub/grub.cfg | nc termbin.com 9999" | 20:14 |
RonaldsMazitis | https://termbin.com/aakt | 20:14 |
EriC^^ | Lustic: try pressing esc to see the background stuff | 20:15 |
RonaldsMazitis | https://termbin.com/1h0ed EriC^^ | 20:15 |
EriC^^ | RonaldsMazitis: type "uname -r" | 20:15 |
RonaldsMazitis | 3.2.0-126-generic-pae | 20:15 |
EriC^^ | RonaldsMazitis: is this a vps? | 20:16 |
RonaldsMazitis | this is my laptop | 20:16 |
EriC^^ | RonaldsMazitis: type "cat /etc/default/grub | nc termbin.com 9999" | 20:16 |
RonaldsMazitis | https://termbin.com/puus | 20:16 |
EriC^^ | RonaldsMazitis: how many hdd do you have installed? | 20:17 |
RonaldsMazitis | 1 | 20:17 |
RonaldsMazitis | but two partitions for two separate linux installations | 20:17 |
EriC^^ | RonaldsMazitis: aha, then the other installation's grub is being used, boot into it and update grub there | 20:17 |
RonaldsMazitis | ah ok | 20:18 |
RonaldsMazitis | thanks | 20:18 |
EriC^^ | np | 20:18 |
RonaldsMazitis | EriC^^ I updated grub in other os | 20:24 |
RonaldsMazitis | still no other kernel in grub menu | 20:24 |
RonaldsMazitis | I just don't want to upgrade from 14.04 on the OS because I like kde 4 | 20:26 |
RonaldsMazitis | I have 18.04 on one partition which is /dev/sda1 | 20:27 |
RonaldsMazitis | and 14.04 on other /dev/sda6 | 20:27 |
RonaldsMazitis | but I can't change kernel in 14.04 | 20:27 |
RonaldsMazitis | it still shows that it is ubuntu 12.04 (version I installed linux on) un additional options only show kernel 3.2 | 20:29 |
RonaldsMazitis | and no other | 20:29 |
RonaldsMazitis | I have installed kernel 4.4 but it does not show in my grub menu | 20:38 |
RonaldsMazitis | EriC^^ what's wrong | 20:39 |
OerHeks | because 14.04 is EOL? | 20:40 |
RonaldsMazitis | ? | 20:40 |
RonaldsMazitis | yeah I know it's old, but I don't like KDE 5 | 20:41 |
TJ- | RonaldsMazitis: "update-grub" should be seen adding the kernel version | 20:41 |
RonaldsMazitis | everything worked until I upgraded to 16.04 | 20:41 |
RonaldsMazitis | does not work TJ- | 20:41 |
RonaldsMazitis | I did update-grub from both OS, | 20:42 |
OerHeks | then you updated the wrong grub? | 20:42 |
RonaldsMazitis | I did update grub from ubuntu 18.04 as well | 20:43 |
TJ- | RonaldsMazitis: "does not work" is no help. when does it not work? what is its error report? with multiple OS installs grub relies on os-prober to discover other OSes so you could try "sudo os-prober" | 20:43 |
OerHeks | so what ubuntu version controls grub? | 20:44 |
RonaldsMazitis | Ubuntu 18.04.3 LTS (18.04) | 20:44 |
RonaldsMazitis | when I sudo os-prober | 20:45 |
RonaldsMazitis | 18.04 | 20:45 |
RonaldsMazitis | I logged in 18.04 updated grub | 20:45 |
RonaldsMazitis | still my other ubuntu 14.04 shows only kernel 3.2 | 20:45 |
RonaldsMazitis | I did not get any errors | 20:46 |
RonaldsMazitis | so | 20:54 |
RonaldsMazitis | what should I do | 20:54 |
tomreyn | RonaldsMazitis: so "sudo os-prober", running on your ubuntu 18.04 installation, returns what? | 20:54 |
RonaldsMazitis | I did not try that | 20:54 |
RonaldsMazitis | I have to restart pc again? | 20:54 |
RonaldsMazitis | does not help | 20:54 |
tomreyn | !enter | 20:55 |
ubottu | Please try and keep as much of your info as possible on ONE line - easier to follow for everyone. | 20:55 |
tomreyn | since we can only support the ubuntu versions listed in /topic, which 14.04 does not belong to, all we can do is see if we can help you make os.prober running on 18.04 detect the kernels off the unsuppoorted ubuntu release you have installed. | 20:56 |
RonaldsMazitis | logically if os-prober returns 18.04 on my 14.04 then it's 18.04, even if it would be 14.04 I tried doing grub-update in both os | 20:56 |
RonaldsMazitis | yeah start nagging me, but I am having 14.04 because I don't like KDE 5 and I have only 27 gb partition for it | 20:56 |
tomreyn | i'm not nagging you, just telling you what you can hope for here and what not. | 20:58 |
tomreyn | i.e. working with os-prober on 14.04 is not something we can support here. | 20:59 |
Guest_8 | Hi can someone help me recover encrypted home dir | 20:59 |
tomreyn | Guest_8: encrypted how? | 20:59 |
Guest_8 | ecryptfs-utils | 20:59 |
tomreyn | which ubuntu version are you running there? | 20:59 |
Guest_8 | 19.04 | 20:59 |
TJ- | Guest_8: did you miss my info earlier? | 21:00 |
Guest_8 | im in a livecd and have installed the utils now | 21:00 |
TJ- | Guest_8: you may also need to do, prior to ecryptfs-recover-private, "sudo keyctl link @u @s" | 21:00 |
Guest_8 | i was in a live cd and i ran out of power | 21:00 |
TJ- | Guest_8: "ecryptfs-recover-private /home/.ecryptfs/$USER/.Private" you mean? | 21:00 |
Guest_8 | what does that command mean | 21:00 |
Guest_8 | I think thats what I need | 21:01 |
TJ- | ^^^ from earlier | 21:01 |
Guest_8 | what does the key ctl commnd do? | 21:01 |
TJ- | Guest_8: fixes a bug in ecryptfs's use of the kernel keyring | 21:01 |
Guest_8 | ok ive run that | 21:01 |
tomreyn | Guest_8: you should also migrate off of ecryptfs | 21:01 |
TJ- | Guest_8: you may not need it, but if you get an error with ecryptfs-recover-private about missing key, use it | 21:01 |
Guest_8 | i am now searching for encrypted directories but it just said permission denied last time | 21:02 |
Gusj | @TJ- Tj, I have installed the new generic kernel with 'sudo apt install linux-generic', have not booted from it yet, received one error "fnfxd.service: Control process exited, code=exited status=1 / fnfxd.service: Failed with result 'exit-code'/ | 21:02 |
TJ- | Gusj: you know where the target system's /home/ directory is? | 21:02 |
Gusj | That was in red, | 21:02 |
TJ- | Gusj: sorry, that was for Guest_8 | 21:02 |
TJ- | Gusj: not sure what fnfxd is | 21:03 |
RonaldsMazitis77 | ok I run os-prober in 18.04 and it shows /dev/sda6:Ubuntu 14.04.6 LTS (14.04):Ubuntu:linux | 21:03 |
RonaldsMazitis77 | I can't get 14.04 to show other kernels | 21:03 |
Guest_8 | im not sure. | 21:03 |
TJ- | Guest_8: have you mounted the target file-system that contains the /home/$USER/ you're trying to decrypt | 21:04 |
Guest_8 | yes | 21:04 |
TJ- | Guest_8: what path have you mounted it on? | 21:04 |
Guest_8 | that is the internal drive in the laptop it is mounts | 21:04 |
Guest_8 | mounted | 21:04 |
Guest_8 | "/media/ubuntu/ebb2d00c-a12e-4207-93a9-f893414b3a4f/home/badger" | 21:05 |
RonaldsMazitis77 | https://paste.ubuntu.com/p/y6jPjTh2tJ/ | 21:06 |
TJ- | Guest_8: that's the user's home directory for a user named "badger" ? | 21:06 |
Guest_8 | yes | 21:06 |
TJ- | Guest_8: so, "sudo ecryptfs-recover-private /media/ubuntu/ebb2d00c-a12e-4207-93a9-f893414b3a4f/home/badger/.Private" should do it | 21:06 |
Gusj | FnFX enables owners of Toshiba laptops to change the LCD brightness, control the internal fan and use the special keys on their keyboard (Fn-x combinations, hotkeys). | 21:07 |
Guest_8 | it returns two lines | 21:07 |
Guest_8 | find: ‘/run/user/999/doc’: Permission denied | 21:07 |
Guest_8 | find: ‘/run/user/999/gvfs’: Permission denied | 21:07 |
Gusj | @TJ- I found that it is a Toshiba ACPI Daemon for Linux | 21:07 |
TJ- | Gusj: if it's an additional DKMS-managed kernel module, maybe it failed for the 4.15 kernel | 21:08 |
TJ- | Gusj: I don't think that'll prevent you rebooting to test the 4.15 kernel though | 21:08 |
TJ- | Gusj: BUT now you're told us about this module, I'm wondering if it is the cause of the lost keyboard/mouse input | 21:09 |
TheFatherMind | Question: I have connected my Amazon Echo Dot as a bluetooth speaker for the computer. But when I use the internal speaker at the same time I get latency on the bluetooth. Is there any way to compensate for this? | 21:09 |
Gusj | @TJ- Ok thank you, I thought now that maybe that has something to do it since with a special FN key one can activate/disable trackpad and probably keyboard it self to lock it? | 21:10 |
tomreyn | RonaldsMazitis77: what does, running on ubuntu 18.04, this return? sudo linuxbootprober /dev/sda6 2>&1 | nc termbin.com 9999 | 21:10 |
TJ- | Gusj: its all about the same keyboard/input devices, so it is the first thing I'd disable in case it is the cause | 21:10 |
Gusj | @TJ- Is ther a way to force activate that module? or should I just reboot and test it to see what is happening? or do I try to disable it now? | 21:11 |
RonaldsMazitis77 | https://termbin.com/3cg9 | 21:11 |
TJ- | Guest_8: so only those 2 errors but no report about "... mounted at /tmp/..." ? | 21:11 |
Guest_8 | no have i not mounted it properly? | 21:11 |
tomreyn | RonaldsMazitis77: what does, running on ubuntu 18.04, this return? sudo linux-boot-prober /dev/sda6 2>&1 | nc termbin.com 9999 | 21:12 |
tomreyn | RonaldsMazitis77: sorry, i had typos in there | 21:12 |
RonaldsMazitis77 | nothing | 21:13 |
TJ- | Gusj: fnfxd isn't a package in Ubuntu; I only find it in Debian | 21:13 |
TJ- | Gusj: correction, found it, I typoed! | 21:14 |
Gusj | @TJ- ah I understand, how could I disable it? | 21:14 |
tomreyn | RonaldsMazitis77: so sudo linux-boot-prober /dev/sda6 returns nothing either? | 21:14 |
RonaldsMazitis77 | tomreyn https://paste.ubuntu.com/p/Yx8Nv7RmV3/ | 21:15 |
TJ- | Gusj: try "sudo systemctl mask fnfxd.service" | 21:15 |
TJ- | Gusj: then try rebooting into the currently broken kernel and see if keyboard then works | 21:16 |
RonaldsMazitis77 | I installed grub-customiser and it shows all kinds of kernels | 21:16 |
Gusj | @TJ- 'Created symlink /etc/systemd/system/fnfxd.service → /dev/null.' | 21:16 |
TJ- | Gusj: if it doesn't then reboot again and try the 4.15 kernel | 21:16 |
RonaldsMazitis77 | for the 14.04 | 21:16 |
TJ- | Gusj: if the keyboard works with the curent kernel it points to fnfxd causing the issue | 21:16 |
Gusj | @TJ- Got it thank you TJ, doing it right now.. | 21:17 |
tomreyn | RonaldsMazitis77: so what you posted now is the output of sudo linux-boot-prober /dev/sda6 right? i'm surprised this didn't get posted to termbin.com, maybe you have this site firewalled. | 21:17 |
RonaldsMazitis77 | I am restarting again , so maybe they show up | 21:18 |
Guest_8 | I have recovered something but it is not my home directory | 21:28 |
tomreyn | so, is it your long lost treasure then? | 21:29 |
tomreyn | is it anything that you recognize? | 21:30 |
Guest_8 | nope | 21:30 |
tomreyn | maybe share the directory structure on a pastebin, if it's not giving away private infomation | 21:31 |
tomreyn | !paste | 21:31 |
ubottu | For posting multi-line texts into the channel, please use https://paste.ubuntu.com | To post !screenshots use https://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. | 21:32 |
Guest_8 | success! | 21:36 |
Guest_8 | ok it is now recovered I think | 21:36 |
Guest_8 | now, how do I make it so that i can do this on reboot | 21:37 |
colimit | tomreyn: I managed to narrow down my problem to the intel_lpss_pci module. If I blacklist that I can boot without pci=nobar and most of the errors that I had before disappear. Didn't solve wireless issues though | 21:38 |
lisa_ | Hello there! I wanted to underclock my system which I did using the cpufrequtils. | 21:41 |
lisa_ | I entered "cpufreq-set ..." in my terminal and my system is running at my desires clock speed. | 21:42 |
lisa_ | However, these change do not persists through reboots. How can I make them permanent? | 21:42 |
sarnold | lisa_: I'd suggest making a systemd unit file to execute your cpufreq commands | 21:44 |
Guest_8 | now that I have accessed my home directory in a live cd... How do I mae this load normally when I reboot? | 21:44 |
lisa_ | sarnold: these commands need to be executed with root privileges which I do by sudo and entering my password. Is that a problem when creating the file? | 21:45 |
sarnold | lisa_: systemd unit files will be executed with root privileges unless you tell it to execute the commands as a specific user | 21:45 |
colimit | tomreyn: Anywaym thanks again for you help. I think the issue is exactly https://bugzilla.kernel.org/show_bug.cgi?id=203485 so I'll go from there | 21:46 |
ubottu | bugzilla.kernel.org bug 203485 in PCI "PCI can't map correct memory resource if the BAR is 64-bit and then leads to system hang during booting up" [Normal,Resolved: invalid] | 21:46 |
glitchd | EriC^^, you busy bud? | 21:46 |
Gusj | @tjsimmons Hi are you TJ- ? | 21:47 |
lisa_ | sarnold: perfect! where can I read some documentation on how to create such a file? | 21:48 |
sarnold | lisa_: I'm trying to find something now :) systemd's insanely complicated and far too flexible, so most documenting I'm finding is vastly too detailed or vastly too basic... sigh | 21:49 |
sarnold | lisa_: here's one of the reference guides .. handy to have around but it's probably not useful immediately https://www.freedesktop.org/software/systemd/man/systemd.unit.html | 21:50 |
sarnold | lisa_: much of this looks decent https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/system_administrators_guide/sect-managing_services_with_systemd-unit_files | 21:51 |
lisa_ | sarnold: the formatting of these files look very complicated; also I can't seem to find any that would be similar to the one I'd want to make and contain bash commands? | 21:55 |
sarnold | lisa_: yeah; sadly it's a lot harder these days to write something simple. the plus side is systemd's reporting and scriptability is much better, but simple things are more difficult now than they used to be | 21:58 |
sarnold | lisa_: I think you'd be setting up a oneshot service | 21:59 |
lisa_ | sooo it would need to contain the line >>ExecStart="cpufreq-set ......"<<? and I also need to set "Before" and "After" myself? | 21:59 |
sarnold | lisa_: the only oneshot services I wrote on my systems run every four hours, which is slightl different than your problem | 21:59 |
=== e is now known as deadk | ||
sarnold | lisa_: ExecStart, yes; I think Before and AFter can be ignored, but you'll need the WantedBy=multi-user.target | 22:01 |
lisa_ | sarnold: what about using /etc/init.d/rc.local? | 22:02 |
sarnold | lisa_: oh man. if systemd still runs that, YES, that'll be way easier | 22:02 |
sarnold | lisa_: heh, searching for that gives this nice example systemd unit file :) https://unix.stackexchange.com/a/479766/7064 | 22:04 |
lisa_ | sarnold: perfect! this is pretty much what I was looking for. Sorry for the stupid questions... I really only know some commands... | 22:07 |
tomreyn | colimit: i'm afraid i lost context. but i'm glad you're making good progress. | 22:07 |
sarnold | lisa_: no worries, that's why people are here :) to help those who ask :) | 22:09 |
lisa_ | sarnold: does it matter where I put my script? on my system there is no /etc/local/libexec/ directory like in the example given | 22:15 |
lisa_ | I meant /usr/local/libexec/** | 22:15 |
sarnold | lisa_: it doesn't really matter; if you really want to execute just that one command cpufreq-something .. then you could put that command in the ExecStart line -- but if you wanted something a bit longer, then putting it in a script does make sense | 22:16 |
lisa_ | well it is 2 commands... one for each core because I couldnt figure out how to give the cpufreq-set command mutiple targets | 22:17 |
sarnold | lisa_: somewhere in /usr/local/ is popular for scripts and programs you put on the system outside of the packaging system -- /usr/local/bin probably exists and would work fine | 22:17 |
lisa_ | would "cpufreq-set 0 --max 2.20GHz && cpufreq-set 1 --max 2.20GHz" work? 0 and 1 being the cores? | 22:18 |
sarnold | lisa_: hmm.. I *think* that you can give two ExecStart= lines in a config and both will be executed. try two lines, ExecStart=cpufreq-set 0 --max 2.20GHz and ExecStart=cpufreq-set 1 --max 2.20GHz | 22:20 |
lisa_ | sarnold: like this? https://paste.ubuntu.com/p/pVc4w9YBr9/ | 22:23 |
sarnold | lisa_: I think no "" marks, none of the ones i see on my system have quotes | 22:24 |
lisa_ | sarnold: alright! saved to the path! I'll reboot and let you know if it works! :) | 22:26 |
sarnold | lisa_: btw you can put the "Untertaktung des prozessors.." bit in a Description= line, so it'll show up in the systemctl status my-service output | 22:26 |
sarnold | lisa_: woot :) goo dluck | 22:26 |
lisa_ | didn't work... double checking my file | 22:32 |
sarnold | lisa_: what does systemctl status my-service report? | 22:33 |
lisa_ | sarnold: Executable path is not absolute: cpufreq-set -c 0 --max 2.2GHz | 22:36 |
lisa_ | so it DOES need a path? | 22:36 |
sarnold | lisa_: ah! so it does! sorry, I never noticed that before. :L( | 22:37 |
lisa_ | no problem sarnold! youre incredibly helpful | 22:38 |
lisa_ | but the textfile I create on the path I specify basically just has to contain the two commands in two lines, right? | 22:39 |
sarnold | lisa_: if you go with the script, you should also start it with #!/bin/sh and set the file executable, too | 22:39 |
sarnold | lisa_: chmod 755 /usr/local/bin/whatever would do the trick | 22:40 |
lisa_ | alright | 22:43 |
lisa_ | it's three lines now, right? | 22:44 |
sarnold | yeah | 22:44 |
lisa_ | Ill go for reboot brb | 22:45 |
lisa_ | still no luck... https://paste.ubuntu.com/p/8pFMfbqHPv/... cpus are at default clock speed | 22:49 |
rfm | lisa_, my guess from the "assignment outside of service" is you put the ExecStart= statement as the very top of the .service file; it needs to be after the [Service] section mark | 22:54 |
lisa_ | rfm: I will try a different solution more along these lines https://forums.linuxmint.com/viewtopic.php?t=203223 | 22:56 |
lisa_ | one sec | 22:56 |
lisa_ | good news! that solution worked like a charm for me. :) | 23:03 |
sarnold | lisa_: was that the /etc/rc.local thing? | 23:04 |
lisa_ | sarnold: nooo, just putting a configuration file for cpufrequtils into /etc/default/ | 23:05 |
sarnold | lisa_: aha! | 23:05 |
lisa_ | thanks so much for your great help though! | 23:06 |
lisa_ | I will have to learn about startup services another day... | 23:06 |
lisa_ | it seems really convoluted to be honest :) | 23:06 |
sarnold | heh, using the things that are already there is probably better in the long run :) | 23:06 |
sarnold | yes | 23:06 |
sarnold | upstart wasn't perfect but it was sure simpler | 23:06 |
lisa_ | yes, i remember there was a lot of commotion in the linux community when they made the switch which was only like 5 (?) years ago? | 23:07 |
lisa_ | but really I don't understand operating systems to that level so I can't give any competent statement | 23:08 |
lisa_ | Have a good night everyone! | 23:08 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!