| === svetlana is now known as Sveta | ||
| === gnomethrower is now known as Wings | ||
| === Wings is now known as wings | ||
| majom | Hi, I have set up googe authenticator authentication for ssh following this guide: https://www.digitalocean.com/community/tutorials/how-to-set-up-multi-factor-authentication-for-ssh-on-ubuntu-16-04. The problem is that according to the logs the google authentication is passes but I still receive a seond prompt to enter my verification code. My logs, configuration and outputs are pasted here: https://paste.ubuntu.com/p/yQZs2XDzn9/ | 08:27 |
|---|---|---|
| aneon | hi, is there any option for encrypted LVM setup in 18.04 install media? | 09:27 |
| majom | aneon, you can choose the encryption method at install time | 09:28 |
| aneon | couldnt find it | 09:29 |
| majom | it's there | 09:30 |
| aneon | under which menu? | 09:30 |
| blackflow | I'm not sure the new subiquity has support for that yet, iirc you'd have to use the old installer | 09:31 |
| aneon | I wont mind old release, I can always do dist-upgrade | 09:32 |
| aneon | the install is failing due to hash mismatch, I tried 6 different mirrors | 09:33 |
| blackflow | it's not old release, but the old installer. there's separate ISOs for that. I think it's called the "Alternative installer" here: https://www.ubuntu.com/download/alternative-downloads#alternate-ubuntu-server-installer | 09:34 |
| aneon | okay I will get that one then | 09:34 |
| blackflow | infact, that installer is the mature one. the new subiquity, that for some reason is now defaulted to, is still in need of baking to completion. | 09:35 |
| aneon | the new thing is horrible | 09:38 |
| aneon | I am getting the alternative installer media | 09:40 |
| aneon | blackflow: Is there any other install media that has all packages bundled in it? | 09:40 |
| lotuspsychje | aneon: feel free to join #ubuntu-discuss | 09:55 |
| aneon | some other time | 09:56 |
| lordievader | The new installer is quite incomplete. I rember I wanted to reuse an existing LVM setup on a host. All the installer could offer me was to create a brand new LVM setup. | 10:42 |
| aneon | dunno why they push beta stuff in release | 10:50 |
| weedmic | anyone familiar with urwid? it's a way to make tui (instead of gui) interfaces? | 12:51 |
| lucido | Hi, anyone here familiar with PAM and google authenticator? | 12:54 |
| chl_ | is anyone using the 'x'-option in tftpd.remap files on 18.04? | 13:03 |
| rbasak | weedmic, lucido, chl_: I suggest you just ask your respective questions. People tend not to volunteer themselves for unknown discussions. | 13:40 |
| teward | lucido: i know a little bit, but i'll need more details on your *question* specifically to help. as rbasak said, you should ask your ACTUAL question otherwise we don't usually volunteer for 'unknown' discussions | 13:50 |
| virole_bridee | Hello. | 13:52 |
| virole_bridee | I have almost no knowledge of Ubuntu Server, and I'm in charge of configuring a new VM that was provided to me. | 13:53 |
| virole_bridee | The guys who created the VM used a 18.04 LTS iso, and I played the installation. | 13:54 |
| teward | > I'm in charge of configuring a new VM that was provided to me. < if you don't have knowledge to do the configuration, then why did they give you the task heh | 13:55 |
| teward | (just saying) | 13:55 |
| virole_bridee | -init | 13:56 |
| virole_bridee | Is Ubuntu server necessarily including cloud-init, or are there iso images without it ? | 13:57 |
| teward | virole_bridee: it's included unfortunately in the Subiqutiy based ISOs. Alternate ISO or mini.iso won't have it, but you can use Cloud-Init to *start* config and install, then remove cloud-init afterwards | 13:58 |
| virole_bridee | as concerns the why : I'm the only Linux guy here, and Ubuntu was forced to us by the final software provider, hence the job for me. | 13:58 |
| teward | which is actually what I do with all my VM(s) deployed from the ISO. | 13:58 |
| teward | cloud-init makes for fast setup and such, but it's lousy after the fact. | 13:58 |
| virole_bridee | OK, thanks. | 13:58 |
| teward | (in standard VM installs anyways) | 13:58 |
| virole_bridee | But it seems to me that removing cloud-init isn't simply done by removing the pakage, right? | 13:59 |
| teward | just `sudo apt remove cloud-init`. It'll remove the underlying service. | 13:59 |
| teward | and then not run anymore. | 13:59 |
| teward | unless you mean something else by 'remove' | 13:59 |
| virole_bridee | Oh, cool. Thanks a lot .Taht's exactly what I was expecting. | 14:00 |
| teward | and yes, just simply removing the package and then doing `sudo apt autoremove` will clean up deps. | 14:00 |
| virole_bridee | some formue | 14:00 |
| teward | it may leave some configuration files behind but they don't do anything with the service removed. | 14:00 |
| virole_bridee | I'll do that | 14:00 |
| virole_bridee | Thanks again, you saved my day. | 14:00 |
| rbasak | teward: I'm curious: how does cloud-init get in the way for you after an instance deployment? | 14:04 |
| teward | rbasak: known issues with hostname not changing, resetting network configs, etc. after the fact. | 14:05 |
| teward | "fixed" but not in 18.04.2 ISOs | 14:05 |
| teward | at least, fixed *supposedly* from what i've heard | 14:06 |
| teward | rbasak: it also irritates the **** out of mee for other reasons, so in *my* case I just purged it from the VM template I use. | 14:06 |
| teward | (since i have a VMware cluster I created a template VM that I just clone now :P) | 14:06 |
| rbasak | teward: but part of the point is that if cloud-init remains, appropriate actions after VM image cloning are automatically taken :) | 14:07 |
| rbasak | teward: are hostname not changing/resetting network configs bugs or just unfortunate defaults for your particular case? Or do you think the defaults are wrong? | 14:07 |
| teward | rbasak: different issue | 14:09 |
| teward | rbasak: the 'defaults' being wrong have issues open | 14:10 |
| teward | there's a current bug in cloud-init where SOMETIMES even if you tell it to not preserve the hostnaame (so you can change hostname on the boxes) it actually *ignores* that setting and keeps resetting hostnames back | 14:10 |
| teward | EVEN IF everything is set to permit hostname changes | 14:10 |
| teward | rbasak: don't get me wrong, I *like* cloud-init | 14:10 |
| teward | but if it's going to be broken when I set it to allow hostname changes and IGNORE that setting... it's on my 'purge after install' list | 14:11 |
| teward | until THAT is fixed... | 14:11 |
| teward | (this bug has been reported also) | 14:11 |
| teward | rbasak: and I've confirmed it in about 20% of my VM deployments personally and at work. The only workaround temporarily was remove cloud-init | 14:12 |
| === cshep is now known as platonical | ||
| rbasak | teward: bug link please? | 14:19 |
| teward | rbasak: gotta dig for it, hang on | 14:20 |
| rbasak | teward: no rush. I'm just curious to track these things :) | 14:20 |
| teward | rbasak: the problem is it's not reproducible all the time | 14:22 |
| teward | it's 'hit or miss' so :/ | 14:22 |
| teward | not sure if the bug has been closed or not already | 14:22 |
| teward | ... fooey and I have a meeting i have to jump into | 14:22 |
| teward | rbasak: I can't find the bug, it may've been closed upstream, but it's a known 'issue' that i've run into multiple times. Might just write a new bug and let someone dupe it to a preexisting bug if there is one | 14:24 |
| teward | ... after this meeting./ | 14:24 |
| virole_bridee | teward: sory for my late answer. 'how does cloud-init get in the way for you after an instance deployment?' : while learning how to configure networking, I arrived at a file under /etc/netplan/ mentionning this cloud-init, | 14:25 |
| teward | virole_bridee: oh, *that* will still exist. | 14:25 |
| virole_bridee | and warning me that all modificaitons would be scratched after reoot. | 14:25 |
| teward | but you can edit that freely | 14:25 |
| teward | and those changes'll stick | 14:25 |
| teward | those only get overwritten when, say, OpenStack or something deploys it | 14:25 |
| virole_bridee | OK. Anyway, I removed this cloud thing, which I do not want at any price on a production server. | 14:26 |
| virole_bridee | NTW, I know quite well Debian/RedHat/CentOS/OpenBSD, I'll have to learn this Ubuntu animal *^v^* | 14:27 |
| rbasak | virole_bridee: most distributions use cloud-init for their cloud images :) | 14:27 |
| rbasak | If you're a professional it's probably worth learning how cloud images work. | 14:28 |
| rbasak | They are very widely used. | 14:28 |
| virole_bridee | Problably, but no cloud here. Purely internal servers. | 14:28 |
| virole_bridee | Also, 'most distributions use cloud-init for their cloud images' : but I don't need a cloud image. | 14:29 |
| rbasak | They are still relevant for internal-only use. | 14:29 |
| virole_bridee | OK, thanks, I'll try to get informed about it. | 14:31 |
| weedmic | I get "python3 is already the newest version (3.5.1-3)" when trying to install/upgrade to 3.6.7 on 4.4.0-1083-aws (ubuntu) - does that make sense/is it telling the truth? | 14:34 |
| rbasak | weedmic: which Ubuntu release? | 14:36 |
| weedmic | Ubuntu 16.04.6 LTS (GNU/Linux 4.4.0-1083-aws x86_64) | 14:36 |
| teward | that's correct for 16.04 | 14:36 |
| teward | weedmic: if you intend to install OTHER python versions beware you might torch your system if you're not careful | 14:37 |
| weedmic | What badness happens if I force it to go to 3.6.7? | 14:37 |
| weedmic | ok, that sounds bad - python is intermixed with the kernel? it's not outside of it and called? | 14:37 |
| teward | it could break anything depending on Py3.5 stuff. If you need 3.6.7 for dev environments I suggest you look into pyenv so you can have 'userspace' installs of 3.6.7 that wont' affect the system packages | 14:37 |
| teward | weedmic: its' not necessarily the *kernel* more like a ton of system utilities and scripts that keep things working | 14:37 |
| teward | (note that pyenv isn't supported in here, but it's my suggestoin) | 14:38 |
| rbasak | weedmic: you might consider upgrading to 18.04. | 14:38 |
| weedmic | can that be done from the kommand line? | 14:38 |
| rbasak | I think most AWS users would expect to deploy a new instance with 18.04 to do what they want. | 14:39 |
| weedmic | i c - i shall resistance meet, but will bring it up. | 14:39 |
| teward | weedmic: i'd back up your stuff first if you want to do an in-place upgrade, but you might do ^ that, deploy a new 18.04, transfer data between old -> new, get things working, decommission old. | 14:39 |
| weedmic | not worried about data - worried about lots of docker containers working and setting up all the face nics and things, but I can script it. | 14:40 |
| rbasak | worried about lots of docker containers working> seems ironic. Isn't not having to worry the entire point of using Docker in the first place? | 14:44 |
| weedmic | all the servers are 18.0x.x except for one - hmmm... | 14:45 |
| weedmic | they talk to eachother and if we leave one link out, poof - hours of work figuring it out. it was assembled by programmers. if "I" did it it would be a script - that is the same each time it is run. | 14:47 |
| ahasenack | rbasak: found this answer for my "what provides <foo>?" question from earlyer | 17:34 |
| ahasenack | aptitude search '~Pdefault-mta' | 17:34 |
| rbasak | ahasenack: nice, good to know thanks | 17:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!