|
I have an Ubuntu Server 16.04 machine where I host a Gitlab instance and play around with various CI/CD tools. It's starting to move from "don't care if it dies and I lose all the data on it" to "would be kind of annoying if it died and I lost all the data on it". The disk is pretty small, so I'd prefer to image it nightly rather than worrying about which folders I need to backup. Is there a simple tool that will take an image of a running system and back it up to a network share? tl;dr: is there a Macrium Reflect or DriveImage XML for Linux?
|
# ¿ Dec 6, 2017 18:38 |
|
|
# ¿ Apr 26, 2024 18:01 |
|
Currently I'm running everything on that server as Docker containers. So yes, I could rsync the volume folders and have all my data, but in case of a screwup of some kind I'd still have to restore everything to the right place, re-start up the right images, etc. And if I run something outside of Docker, I'd need to remember to add the right extra stuff to the backup. Since the entire system is using a little over 100GB in disk space, I'd much prefer to image the whole thing and restore it with one click.
|
# ¿ Dec 7, 2017 00:31 |
|
I've got a pretty n00b question. I'm reasonably familiar with cloud stuff but I'm very lacking in the meat-and-potatoes linux admin stuff. I bought a Raspberry Pi 4B which I want to use as a headless docker server for Bitwarden, Nextcloud, etc. I initially installed Raspbian Lite and it worked great, but I wanted a 64-bit image, and eventually settled on this one. (I'm still reeling from the idea that in 2020 the best documented, most idiot-proof, most plug-and-play distro I could find was a gentoo one.) It works great except for one thing, not at all essential but I'm curious. When I booted up Raspbian Lite and it auto-connected to the wifi, I could immediately SSH into it by hostname (ssh pi@raspberrypi). The Gentoo image does not do this, even though it appears in my home router with a proper hostname and all... and even though I have the hostname set up in the /etc files (/etc/hosts was already there, I added /etc/hostname myself and then rebooted) ... I can only SSH to it @192.168.1.62, my (Windows) desktop won't recognize @pi64. In particular, ping -a 192.168.1.62 does not display the hostname. What could this Gentoo image not be doing, that Raspbian used to do?
|
# ¿ Apr 28, 2020 22:33 |
|
Thanks for the hints. I looked up avahi and it was what Raspbian was using, so I'm gonna go with that one.
NihilCredo fucked around with this message at 10:44 on Apr 29, 2020 |
# ¿ Apr 29, 2020 08:14 |
|
Ok, it *almost* works. It correctly broadcasts its name, but SSH doesn't seem to like it and resets the connection after password verification. Any clues where I could look for logs into this error? edit: I've discovered the -vvvvv option and i'm trying now to look up the last message (client_check_window_change) and the error codes at the end, with little succes EDIT2: Ok, it's not the window change (that's just from resizing the terminal). I've tried a bunch of times and it fails at various points after the login, but always with the line recv - from CB ERROR:10060. EDIT3: It's a problem with the Windows OpenSSH client, if I connect to root@pi64 from Linux it works. Nevermind. NihilCredo fucked around with this message at 11:34 on Apr 29, 2020 |
# ¿ Apr 29, 2020 10:45 |
|
Mr Shiny Pants posted:Try Putty? No need, I can just use WSL. Thanks anyway.
|
# ¿ Apr 29, 2020 14:26 |
|
I tried VFIO, got it to work after more than a little troubleshooting and blind "e ctrl-x" grub editing, and realized it was much clunkier than I thought. What I thought it was: I can use Linux normally, but when I start a VM image I can temporarily give it control of the GPU and its attached monitor, same as any USB device. So I can use Linux all the time, including for most gaming, and fire up a Windows VM only when I want to play a game that won't run on Lutris or a Windows-only application. If I have a secondary screen and GPU, I can keep using Linux on them while Windows is running on the main. What it apparently is: the main GPU is permanently assigned to the VFIO driver and can ONLY be used from the VM. You can't use the GPU from Linux at all, all your gaming must be via VM. You can't use your main screen outside a VM, not even for booting up, unless you change its input to the secondary GPU (manually or with a hardware HDMI switch). The secondary screen should actually work as intended, although I don't have room for a second screen right now so I couldn't test it. Have I got this right this time?
|
# ¿ May 31, 2020 10:05 |
|
A properly configured reverse HTTP proxy should send the X-Forwarded-For header containing the original client IP. You should then configure your application logs to register that: https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/X-Forwarded-For
|
# ¿ Jun 17, 2020 06:49 |
|
Paul MaudDib posted:Google never deletes anything, but the backups are all encrypted with per-user encryption keys, so they can delete the key and the backup is just so much random noise. That seems like a good approach. I mean, if you have per-user backups the whole problem doesn't apply in the first place, since you could just delete them along with the data even if they weren't encrypted. It's a problem for multi-tenant companies that don't have per-user backup.
|
# ¿ Jul 5, 2020 09:13 |
|
apropos man posted:e: To give an extra shoutout to youtube-dl. I've been using that for ages to grab TV shows and put them on my jellyfin server. Do you just run youtube-dl manually from the server's terminal, or do you do it through some kind of frontend?
|
# ¿ Jul 12, 2020 15:56 |
|
I would just install a music manager / streaming server directly on the NAS, if at all possible. Ampache or Airsonic or a bunch of others.
|
# ¿ Jul 30, 2020 12:41 |
|
hifi posted:My other suggestion would be, most web apps let you set a HTTP root, so any url you click in the app can be prefixed with "http://butt.com/app1" for example, so you can host multiple things on the same domain without using extra ports or subdomains. Subdomains are the way to go, as long as you have a halfway decent domain registrar that will let you register a wildcard subdomain. Then you just publish each service on service.mydomain.com, let your proxy register a Let's Encrypt certificate for each one (or a wildcard certificate), and you never need to even think about whether your services are properly dealing with custom root paths.
|
# ¿ Aug 1, 2020 11:05 |
|
RFC2324 posted:Since when does the registrar need to allow subdomains? If you register company.com, do you no longer control *.company.com? Yeah if you have everything on one IP it's not a problem. But it's not uncommon that company.com is your public/marketing landing page and it's on a totally different infrastructure than your internal services like jira.company.com, erp.company.com, etc. Of course, you can also just put those on company-services.com or whatever.
|
# ¿ Aug 1, 2020 16:27 |
|
astral posted:Actually, I re-read it and have to revise my previous post: they meant being able to set things up so DNS would respond to anything.domain.com with a certain IP. Correct. The web UI of the registrar I personally use lets me use asterisks in DNS records, but the one my company uses doesn't, so in order to deploy a new app we need to have someone define records for app.company.com. It takes like a minute so I haven't pushed to change registrar over it, but it's still totally unnecessary.
|
# ¿ Aug 2, 2020 12:21 |
|
You don't need to tell me to get a better provider. I already have a good one for myself, the "no wildcards allowed" issue is only with the one my company uses (and it's nowhere close to worth making a fuss over, or even finding a dedicated provider separate from the domain registrar + web host), and in fact the whole reason we are having this discussion in the first place is because I mentioned to another goon that they should check if their provider supports wildcards - which I thought was a given until I discovered one that didn't. Let's move on.
|
# ¿ Aug 2, 2020 22:04 |
|
wormil posted:Also, as an aside if anyone is interested. I've tested 7 distros on this old laptop and Peach/Patriot is the snappiest, not measured, just based on use. Peppermint and Mint are very close seconds, Pop! is a very close third. Any of those four are fast enough for basic stuff. Ubuntu and Minjaro were faster than Windows but sluggish and would be mildly annoying in use unless you've never used anything better. Also tested Chromium but that was before I fixed the laptop overheating so it's not fair to comment but I believe it would be fine if you want a google ecosystem. Did you choose XFCE for all of those distros? Otherwise it's not surprising that an XFCE distro (Peach / Patriot) feels faster than those that default to GNOME.
|
# ¿ Sep 28, 2020 12:35 |
|
D. Ebdrup posted:In case wayland is a thing anyone's interested in, there's a 3D compositor based on wlroots, inspired by compiz - and it as well as its plugins also available on FreeBSD. Do any of the more popular compositors / desktop environments support something like this tiling plugin? (Tiling WMs not included, of course - tiling windows is something I find occasionally useful but do not want to always have active) I really like the way this Wayfire implementation looks, it seems better than Windows's: https://www.youtube.com/watch?v=SKfhPsA8TxU
|
# ¿ Oct 13, 2020 17:54 |
|
Is Clonezilla still the simplest option to backup a running system's OS drive against hardware failure? User data would be handled separately.
|
# ¿ Oct 14, 2020 18:25 |
|
Methanar posted:rsync -a --backup --backup-dir=backup /dev/sda1 destination/ That won't cover boot stuff () like grub and such though, right? So if the OS drive died I would have to install a fresh linux and then rsync back everything on top (will that work on a running system?).
|
# ¿ Oct 14, 2020 19:58 |
|
xtal posted:You should probably do that anyway like the person above you said. Just copying the boot folder might not work either, for example if you have that on an EFI boot partition. Maybe... To add some more details, the use case is that I have a Pi 4 which is notorious for killing SD cards. While all the actual user data is on external storage, and the various configuration files are backed up, I think it's now worth my time to take a full image of the SD card, so if it fails I can just burn the last image on a new card and bring it back online, instead of having to reinstall the OS and manually reapply all the various configurations.
|
# ¿ Oct 14, 2020 21:13 |
|
xtal posted:Now is a good excuse to learn Ansible I skipped the era of configuration management and moved on straight to containers, I'd rather not have to learn it all just for a homelab box, thanks minato posted:podman is a dropin replacement for docker, in most respects. The main external difference is that it doesn't need to run a root daemon. It also lacks docker-compose. There's a project to write a drop-in replacement but it's still under development and very far from production ready. That's important because there's a lot of space between "simple enough that I can launch individual 'podman run' commands manually" and "complex enough that I must dehumanize myself and face tp K8S".
|
# ¿ Oct 14, 2020 22:32 |
|
Mr Shiny Pants posted:PXE it? Not a bad option, but the only candidate for a PXE server in my house is my gaming desktop and I wouldn't rely on it.
|
# ¿ Oct 15, 2020 09:10 |
|
Bob Morales posted:Random-rear end question, somewhat related to Linux: I have a single large USB drive with both the latest win10 ISO and about a dozen different linux ISO, all of them bootable at the same time. Ventoy is loving amazing. Here's the user guide: 1) Install Ventoy on a USB drive 2) Copy however many ISO you have in the main folder 3) Boot from the drive and Ventoy will let you pick which ISO you want to boot from
|
# ¿ Oct 28, 2020 21:16 |
|
I'm getting this error on WSL Debian when VS Code tries to install some plugin:code:
I thought that ~ constraint meant " anything under 8.4.0-2 or anything that's at most minor bump ahead" , am I wrong?
|
# ¿ Nov 11, 2020 12:31 |
|
G604 is the perfect mouse as far as I'm concerned. "Only" six thumbs button but they're actually usable while holding your thumb in a comfortable position, plus wireless/bluetooth and metal toggleable scroll wheel. The Linux keymapper app is also really good and kicks rear end over the official Windows piece of poo poo. That said, nothing made in these fallen days will again equal the glory of the OpenOffice®* Mouse: https://www.geeky-gadgets.com/openoffice-mouse-10-11-2009/ * it was not actually affiliated with openoffice in any way, the maker just slapped their name on it and hoped they wouldn't notice or care (they did and they did) NihilCredo fucked around with this message at 02:08 on Nov 23, 2020 |
# ¿ Nov 23, 2020 02:05 |
|
they totally deserve poo poo, at least a massive loss of trust, for cutting centos 8's EOL by several years over what was promised but idk, I don't really feel the outrage for "enterprise software company will no longer spend work on a free clone of its main product". what's the Venn diagram intersection of - must absolutely have all the latest security updates ( = can't just stop updating centos) - cannot take even the smallest risk of an upgrading issue ( = can't update centos in production) - can't take the time to properly test centos updates in a staging environment - can't afford to pay for RHEL ? like, if you're _not_ running a business, just update and worst case scenario you'll have a bit of downtime if some library has a bug. you do take backups, right? if you _are_ running a business, then I don't really think you're entitled to have someone else maintain your business-critical os for free. asking you to try out new releases in a staging env in exchange for maintenance is imo more than fair, even when it's IBM asking that idk it reminds me of companies becoming outraged that random github projects aren't being kept up to date when they haven't ever contributed a dollar or a pull request
|
# ¿ Dec 9, 2020 14:26 |
|
I just remembered that Fedora Server exists. Who are its target users, exactly? People who run pet servers and can afford to update them both carefully and frequently? People who for some reason need to keep package versions synchronized between their workstation(s) and their server(s)?
|
# ¿ Dec 16, 2020 17:28 |
|
Haven't most of the anti-systemd crowd consolidated behind openrc which is a few years older than systemd anyway Also, is openrc a horror? Genuine question, I have no clue, it's just my default assumption
|
# ¿ Jan 3, 2021 00:06 |
|
I think that's why Garuda is getting so much hype, it's Arch/Manjaro with Timeshift installed by default
|
# ¿ Jan 26, 2021 10:51 |
|
Nitrousoxide posted:Neato, my entire Garuda install nuked itself overnight, including the snapshots. Won't even load into the GRUB bootloader without complaining. Christ on a bike. Can you please let us know if it turns out it was a HDD death rattle or something else? Because it makes me want to stay the hell away from Garuda the next time I install a linux, but I'm basically the target user so it'd be a pity to discard it over a false alarm.
|
# ¿ Jan 26, 2021 17:07 |
|
Mr. Crow posted:That's actually pretty funny the UI accounts for that. Reminded me of https://m.imgur.com/gallery/JADYkVb
|
# ¿ Jan 31, 2021 20:51 |
|
Nitrousoxide posted:I'm really appreciating how easy it is to create a bootable linux distro USB stick in any OS now. I just had the unfortunate experience of trying to create a windows bootable USB without any windows machines. Use Ventoy for everything. Dump the windows .iso of your choice in the main folder and It. Just. Works. I have had a single boot USB for a while now with whatever the latest release of win10 is and a bunch of Linux distros, all together.
|
# ¿ Feb 7, 2021 13:25 |
|
xtal posted:Part of the reason for this is that btrfs and zfs are part filesystem and part volume manager. With ext4 you get extra features from layering it with other software, like lvm RAID and LUKS encryption. Am I reading this right, that ZFS and BTRFS do not work well (or at all) with LUKS? Because at-rest full-disk encryption seems extremely important for laptops which can get stolen quite easily.
|
# ¿ Feb 26, 2021 22:23 |
|
Antigravitas posted:It does have the best options though. This just gave me a freedom boner.
|
# ¿ Mar 10, 2021 21:48 |
|
DerekSmartymans posted:Are there any distros that work with WSL2 that aren’t on the Microsoft Store? I’d never heard of it before and it’s much easier than setting up a VM, as well as tons of setups on YouTube and tutorials (I haven’t set up a VM since around Ubuntu 12? 13? Long ago). I’m sure the process is seamless and faster these days, but I was bringing up a bash terminal in WSL2 Debian within like 10 minutes. It’s more fun to play around and crash doing poo poo and learning not to do things than dual booting, too 🤓. Gentoo runs on WSL2 and it's pretty much the ultimate "play around and crash doing poo poo and learning not to do things" distro.
|
# ¿ Apr 1, 2021 17:45 |
|
Also check out jellyfin instead of plex. The LinuxServer guys have made a good image with good documentation in case you need stuff like 4K hardware acceleration: https://docs.linuxserver.io/images/docker-jellyfin
|
# ¿ Apr 8, 2021 09:40 |
|
Mr. Crow posted:Please don't use anything by LinuxServer if you plan on exposing it to the world at all, at least a couple years ago it was a complete poo poo show. Can you expand on the poo poo show? I normally stick to the official images whenever possible, but for Jellyfin in particular I have in fact been using the linuxserver image because the official one had trouble on the Pi (don't remember exactly what).
|
# ¿ Apr 8, 2021 22:47 |
|
Mr. Crow posted:I haven't looked at it recently but I looked at it back in the day for plex and transmission or something and it was basically just running everything as root, full CAPS, pull as many dependencies into the image as possible and just... The antithesis of a discrete unit of processing containers are supposed to be I see. Well I started using it sometime mid-2020, and they go out of their way to not have you run their images as root, as for caps, I think those need to be explicitly enabled by the user with --cap-add, which they don't seem to require or suggest. Can't speak about dependencies in general, but here's the Jellyfin dockerfile and I'm not seeing anything that shouldn't be there. I don't doubt you, it's quite possible they improved their practice since its origins as (IIRC) basically a hobby project.
|
# ¿ Apr 9, 2021 09:11 |
|
Gentoo noob question! I've got a headless Pi running gentoo, and I needed to extract a .7z archive, so I tried installing app-arch/p7zip. From what the wiki says, the package includes a GUI only if the wxwidgets USE flag is enabled. Here's the output of emerge --ask --tree --verbose. It looks like it wants to install a bunch of stuff including Wayland: I tried checking with euse -i wxwdidgets, and sure enough, the local flag is in state 'D', apparently required by some Qt ebook reader I absolutely don't want or need: code:
code:
code:
|
# ¿ Jun 28, 2021 16:52 |
|
|
# ¿ Apr 26, 2024 18:01 |
|
Antigravitas posted:So, echo "app-arch/p7zip -wxwidgets" >> /etc/portage/package.use/p7zip should tell portage to not pull in wxwidgets for that package. If /etc/portage/package.use doesn't exist, create it. Oh, I didn't know that was an option! I thought that I had to get rid of the USE flags, and since I had that coolreader package installed, I needed to first uninstall it before I could remove the flag. Thanks, it worked perfectly. Also the euse output I quoted didn't mean that coolreader was installed?
|
# ¿ Jun 30, 2021 15:20 |