Beret | å/b 5 | 02:43 |
---|---|---|
bigdaddy63376 | My laptop suspended when I closed the lid. I think a server OS should not do that. I would like /etc/systemd/logind.conf in the install media to have HandleLidSwitch=ignore | 05:49 |
bigdaddy63376 | Do I need to do this myself and get access to place a pull request? | 05:59 |
bigdaddy63376 | Or is there someone here who knows about the maintainer for the package that contains that file? | 05:59 |
mybalzitch | you can edit the file yourself | 06:05 |
Maik | bigdaddy63376: i'd go for a raspberry pi 4 and uses that as a server, laptop aren't really made for that. | 07:42 |
znf | if you ever manage to find a rpi in stock :) | 07:59 |
ahasenack | is there some sort of "who am I" command for the current ubuntu release I'm running? | 13:32 |
ahasenack | or should I just parse /etc/os-release? | 13:32 |
ahasenack | I didn't want to source that whole file for just one variable | 13:32 |
ahasenack | ah, there is lsb_release -cs | 13:33 |
jfsimon1981_b | Hi, Could someone let me know how we properly do a subnet with nat ? I have a lan at 10.0.*.* and i need to create a subnet 192.168.100.* with nat and dhcp for test purposes | 13:38 |
jfsimon1981_b | https://pastebin.com/dKFSqUHm | 13:38 |
jfsimon1981_b | I crafted a configuration file though it's not yet working, this is the complete file | 13:39 |
jfsimon1981_b | Thanks | 13:39 |
znf | ahasenack, the preferred and reliable way is to source /etc/os-release | 14:19 |
ahasenack | I don't want to potentially "pollute" my shell script with random variables | 14:19 |
znf | you're not poluting with "random" variables | 14:19 |
ahasenack | well, it's an external file I don't control | 14:20 |
znf | but that's the proper and reliable way to determine OS across multiple distributions and linux releases | 14:20 |
znf | has been a standard for a few years now | 14:20 |
znf | a minimized 22.04 doesn't have lsb_release, AFAIK | 14:20 |
sdeziel | `awk -F= '/^VERSION_CODENAME=/ {print $2}' /etc/os-release` should avoid the pollution | 14:22 |
znf | I'm sure the 1kb of memory saved will be appreciated :P | 14:26 |
samy1028c | Hello all. We have a system with exim4 running with custom user:group. However, every so often (6 or more months) it seems something on this Ubuntu 20.04 FIPS LTS is resetting permissions back to Debian-exim:adm. Any ideas what to look at? | 15:34 |
samy1028c | We updated the cron.daily/exim4 to have the custom user:group when tidydb runs. However, even if something daily ran, it doesn't explain why after 9 months the permissions reset. | 15:36 |
samy1028c | Is there anything in Ubuntu itself that might be causing this? | 15:36 |
sdeziel | samy1028c: does the permission/ownership reset happens at the same time the exim4 package is updated? | 15:38 |
samy1028c | sdeziel: ahh, actually, I hadn't looked at dpkg.log. It does show that exim4-base updated at the time exim4 stopped working. | 15:40 |
samy1028c | hmm, so a question - how do I get it to not clobber the permissions? | 15:40 |
rbasak | https://git.launchpad.net/ubuntu/+source/exim4/tree/debian/exim4-base.postinst?h=applied/ubuntu/focal-devel resets the permissions I think? | 15:40 |
rbasak | I'm not sure the packaging supports use of a custom user/group. | 15:40 |
samy1028c | so, why did it even update? We purposefully do not install updates automatically because we have to test through our scripts and such to ensure non-breakage. | 15:41 |
rbasak | How have you configured updates not to install automatically? | 15:42 |
samy1028c | I run "apt list --upgradeable" and there are a number of items waiting on us to trigger the updates on after regression testing/QA is completed on our dev systems. | 15:42 |
samy1028c | What exactly does "install" do in init.d? ./rc3.d/S01exim4: install -d -oDebian-exim -gDebian-exim -m750 /run/exim4 | 15:43 |
samy1028c | rbasak, how can I double-check the automatic updates? | 15:43 |
samy1028c | ahh, /etc/apt/apt.conf.d/50unattended-upgrades | 15:45 |
rbasak | It's pretty hacky/unsafe to rely on not installing updates to avoid an update from reverting an unsupported configuration change. I'd find another way to solve your problem. | 15:46 |
samy1028c | yeah, perhaps a cronjob to check permissions regularly. | 15:50 |
samy1028c | We may also review our need for the custom permissions as this may have been required when we first deployed this system on 18.04 several years ago. | 15:50 |
rbasak | No, that's just as bad. Stop chasing your tail! | 15:50 |
bG9s | I mean restorecon exists to basically check contexts (basically similar to permissions) | 15:56 |
bG9s | not a bad idea to make sure permissions and contexts are in order | 15:57 |
=== elastic_dog is now known as Guest5195 | ||
ahasenack | can somebody see a mistake here? It's like "apt-get update" wasn't run. If I shell into the container, run apt-get update, then try to install those strongswan packages, it works: https://pastebin.ubuntu.com/p/Nx9JwV8gn7/ | 17:56 |
patdk-lap | you are even running cron inside the container? | 18:00 |
patdk-lap | none of mine do, so it would never autoupdate a container | 18:00 |
=== blackboxsw_away is now known as blackboxsw | ||
ahasenack | found it, cloud-init was still running in the background and changing sources.list, which invalidated the first apt-get update run | 18:32 |
=== lord_black is now known as lord_daemon | ||
sarnold | ahasenack: that feels like it's worth a bug report, perhaps cloud-init ought to be locking the apt frontend lock or something while doing its thing | 20:45 |
ahasenack | there is this trick that I have seen other use, and I am using now too: | 20:46 |
ahasenack | lxc exec "${container}" -- cloud-init status --wait | 20:46 |
sarnold | oh nice. i mean, a bit annoying, but at least there's a thing for that :) | 20:47 |
blackboxsw | +1 on that blocking `cloud-init status --wait` trick (sry was afk earlier). But, if one process is running something like ` DEBIAN_FRONTEND=noninteractive apt-get install` shouldn't the apt/dpkg locks be setup already to block other services trying to install pkgs? | 20:49 |
ahasenack | that is not the issue | 20:49 |
ahasenack | sources.list was rewritten after my first apt-get update | 20:50 |
ahasenack | so that invalidated that apt-get update | 20:50 |
ahasenack | it was like this | 20:50 |
ahasenack | apt-get update | 20:50 |
blackboxsw | ahh ok | 20:50 |
ahasenack | cloud-init rewrites sources.list | 20:50 |
ahasenack | apt-get install <package> | 20:50 |
ahasenack | package: not found | 20:50 |
blackboxsw | ahh so race on config file update vs final apt-get install | 20:50 |
ahasenack | yeah, I was doing "apt-get update && apt-get install package" | 20:51 |
ahasenack | but in between, sources.list got rewritten by cloud-init | 20:51 |
blackboxsw | yeah, hard to avoid that situation without either providing the directives in #cloud-config specifically via `packages: [package]\npackage_update: true` or by just externally calling cloud-init status --wait (or setting up a systemd service that is `After=cloud-init.target` | 20:56 |
k8yun | hello... does anyone have a good tutorial link to setting up postfix in ubuntu? Looking to only send emails out as notification system. | 23:38 |
sarnold | k8yun: try https://ubuntu.com/server/docs/mail-postfix | 23:40 |
Teridon | ubuntu autoinstall question with cloud-init ssh_keys. During autoinstall, I see that /etc/ssh/ has the keys I specified in my user-data file. But after install finishes and reboots, it has new keys. I do have "ssh_deletekeys: false" in the top level of the user-data. It's also generating a dsa key, despite the specified "ssh_genkeytypes" https://dpaste.org/q1dPA not sure if my PXE boot line is relevant; the "cloud-init-bios | 23:40 |
Teridon | " directory has the user-data file https://dpaste.org/RWVBP | 23:40 |
k8yun | thank you. I will go through the link. I had gone through a few links and all of them did not allow me to send an email out so far but I will give this link a try. | 23:41 |
sarnold | k8yun: depending upon where you're hosting your machine, itmight be worth talking with your isp or managing 'security groups' on the machine or whatever | 23:42 |
sarnold | k8yun: it's pretty common for smtp to be blocked everywhere to cut down on spam | 23:42 |
sarnold | k8yun: .. and also pretty common for smtp recipients to be brutal with RBLs and so on to block mail from dynamic ranges, entire countries, etc | 23:43 |
k8yun | I see. I am not too familiar with how smtp works. I tried to send to my work email then the `/var/log/mail.log` showed that it was trying to connect to my work domain. | 23:43 |
k8yun | so I think I might need to implement some security measure in the set up so that my work email server recognizes it as a safe sender or something... not sure... I will go through the link provided and find out more I guess | 23:44 |
Teridon | related: if your DNS doesn't have SPF setup you're probably also going to have trouble | 23:44 |
Teridon | is this channel archived? if so where? my VPN is sometimes unstable and I don't want to miss a reply | 23:46 |
blackboxsw | Teridon: archive of this channel https://irclogs.ubuntu.com/2022/11/28/%23ubuntu-server.html it's a few hours behind | 23:50 |
Teridon | ty | 23:50 |
blackboxsw | Teridon: also autoinstall question. the Ubuntu live installer (server and desktop) actually runs cloud-init in two boot stages. if you provide top-level #cloud-config keys outside of 'autoinstall:' those keys will apply the the ephemeral pre-provisioning environment (not the target installed system after reboot) | 23:51 |
blackboxsw | if your keys you pasted are provided below autoinstall:user-data:X,Y,Z then they'll apply to the final installed environment across reboot. | 23:52 |
sarnold | blackboxsw: that sounds like a very useful thing to know. do you know if it's written down anywhere? :) | 23:57 |
sarnold | aha! it is! it is! though I think I might have overlooked it if I hadn't been looking for it | 23:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!