Anaconda-devel

anaconda-devel@lists.stg.fedoraproject.org

6518 discussions

Re: auditd replacement in installer environment
by Brian C. Lane 15 Sep '22

15 Sep '22

On Thu, Sep 15, 2022 at 10:20:04AM -0400, Steve Grubb wrote: > Hello, > > On Thursday, September 15, 2022 8:11:39 AM EDT Vladimir Slavik wrote: > > while trying to make sense of the remaining C bits in anaconda, I found > > that we actually have a mock auditd, which does nothing and replaces the > > real auditd on boot.iso, via lorax templates. > > > > Now I'm trying to understand why. Is it because it writes too much to > > journal? Is it because it takes 90 MB memory? Something else? > > > > Steve, Brian - would you know? > > > > PS: https://github.com/rhinstaller/anaconda/pull/4331 - moving it from the > > python module directory where it was hiding. > > I would guess that they are trying to prevent hardwired audit events from > going into the install logs. If you boot with audit=0, you wouldn't need a > mock auditd because auditing is disabled...except that systemd-journald > blindly enables auditing. Maybe they fixed it to respect the command line by > now, I don't know. > > Another item, and maybe this is the reason, if there is no auditd, selinux > sends AVC's to syslog. So, maybe it's to suppress AVC's? > > I'd suggest booting with audit=0. If you get any events in your logs, you can > probably replace auditd with a python variant. Libaudit has python bindings. > It is not well tested for handling audit events. But it is used by semanage > and some other python programs. I talked to pjones (update in the PR as well) and it was because space was limited and something needed to listen to the socket to throw away the messages. My guess is that this is still valid, I don't think it would provide anything useful in the installer environment, and would just clutter up the journal with irrelevant messages. Maybe. Booting with audit=0 may work, but one thing to remember is some people use PXE and wouldn't have the same cmdline as the iso, so it may be better to just keep this so it 'Just Works(TM)'. Brian -- Brian C. Lane (PST8PDT) - weldr.io - lorax - parted - pykickstart

1 0

auditd replacement in installer environment
by Vladimir Slavik 15 Sep '22

15 Sep '22

Hello, while trying to make sense of the remaining C bits in anaconda, I found that we actually have a mock auditd, which does nothing and replaces the real auditd on boot.iso, via lorax templates. Now I'm trying to understand why. Is it because it writes too much to journal? Is it because it takes 90 MB memory? Something else? Steve, Brian - would you know? PS: https://github.com/rhinstaller/anaconda/pull/4331 - moving it from the python module directory where it was hiding. Best, Vladimir -- Vladimír Slávik <vslavik(a)redhat.com> Software Engineer, Platform Engineering Red Hat Czech, s.r.o.

1 0

lorax: fstrim by default for image installs
by Chris Murphy 22 Aug '22

22 Aug '22

Hi, I'm trying to track down why image installs still aren't being trimmed. It was implemented in lorax, and also the buildvm's in Fedora infra are using qemu discard=unmap, so it should just work (it works locally with virtio block devices, the virt-manager default). Workstation Live x86 koji logs includes https://kojipkgs.fedoraproject.org//packages/Fedora-Workstation-Live/Rawhid… Which shows 2022-07-22 00:41:33,905 INFO: Running... fstrim -v /mnt/sysimage/ 2022-07-22 00:41:34,113 INFO: /mnt/sysimage/: 2.1 GiB (2205618176 bytes) trimmed 2022-07-22 00:41:34,113 DEBUG: Return code: 0 That's expected. However, it's no smaller than several weeks ago when discards weren't passing through from guest buildervm to the host with the sparse file. I'm still pondering how to track down how we can get a return code of 0, and yet the trim didn't actually do anything. Next, two examples of oz logs: https://kojipkgs.fedoraproject.org//packages/Fedora-Cloud-Base/Rawhide/2022… https://kojipkgs.fedoraproject.org//packages/Fedora-Workstation/Rawhide/202… But fstrim isn't mentioned, so I'm suspicious oz isn't including all the anaconda logs, which would be an oz bug. -- Chris Murphy

3 7

Drop the protocol selection in the Installation Source spoke
by Vendula Poncova 09 Aug '22

09 Aug '22

Hi! We are planning an extensive clean-up of the code for the Installation Source spoke to prepare the code for the payload modularization. That creates a good opportunity for revisiting the design of the user interface as well. We would like to address some of the long-term issues and propose an alternative. It shouldn't be anything drastic, just a small reorganization of the current components and options. See the proposal at https://github.com/rhinstaller/anaconda/discussions/4257 for more information. Vendy

1 0

Making Anaconda do hybrid BIOS+UEFI setups for legacy BIOS installs
by Neal Gompa 18 May '22

18 May '22

Hey all, As y'all may know, there was the rather contentious discussion around deprecating/removing legacy BIOS support (which was ultimately rejected). As part of that discussion, a more incremental path to deprecating BIOS boot was identified. The first step of that is to move to legacy BIOS installations to GPT partitioning[1][2]. Following that, I'd like to adjust Anaconda to install a hybrid BIOS+UEFI boot setup for BIOS systems, similar to what I did for Fedora Cloud[3]. However, I'm not really sure how to implement that and would like some guidance on how to pull it off. Thanks in advance and best regards, Neal [1]: https://fedoraproject.org/wiki/Changes/GPTforBIOSbyDefault [2]: https://github.com/rhinstaller/anaconda/pull/4104 [3]: https://fedoraproject.org/wiki/Changes/FedoraCloudHybridBoot -- 真実はいつも一つ！/ Always, there's only one truth!

4 6

Feedback on new Anaconda Insatller
by Florencia Forno 16 May '22

16 May '22

Hi everyone, The User Experience Research team is conducting a study on Anaconda Installer's new interface. If you are interested in an opportunity to influence the future of Anaconda Installer, please fill out this brief survey <https://redhatdg.co1.qualtrics.com/jfe/form/SV_3wVqDVLem1HPyK2>. Browser link: https://redhatdg.co1.qualtrics.com/jfe/form/SV_3wVqDVLem1HPyK2 We will contact you if your experience matches what we need for this study. We expect to begin in early June. Thank you, *Flo Forno* UX Researcher *Dallas, TX* --

1 0

Silverblue & Kinoite to use separate "var" subvolume by default
by Chris Murphy 07 Feb '22

07 Feb '22

Silverblue users have occasionally expressed a desire to snapshot the "root" subvolume separately from /var so they can rollback "root" separately from /var content. Since rpm-ostree puts the rpmdb in /usr, and Btrfs subvolumes have essentially no cost, it's a nice tweak for these users. So I've put together a self-contained change proposal to do that. https://fedoraproject.org/wiki/Changes/VarSubvol4SilverblueKinoite I've done quite a few such installations already and the installer has no issue with it, and systemd fstab generator reliably mounts "var" to /var early enough that there's trouble there. I consistently see the journal flush to persistent storage *after* /var has mounted which is what we need. But if anyone can think of additional liabilities, let me know. A possible gotcha I'm not certain about, is if the Btrfs autopart scheme is modified to add a /var mount point, we need to make sure it only applies to Silverblue and Kinoite. And only applies to the Btrfs scheme, not LVM or Standard partitioning schemes. If that's too tricky in the Fedora 36 time frame, then we'll just punt. Thanks! -- Chris Murphy

5 7

partitioning, Everything ISO
by Chris Murphy 25 Jan '22

25 Jan '22

cc: anaconda-devel@ From today's Fedora QA minutes https://meetbot.fedoraproject.org/fedora-meeting/2022-01-24/fedora-qa.2022-… 16:20:54 <pboy> I'll test that again, but I'm sure, if you use everything DVD you get the wrong preconfiguration. We had a thread about that on server list. I'm not finding this thread. But Adamw surely knows the history of Everything ISO better than I do. My fuzzy memory is that it's sort of a side effect of how images are made and it's kept around as a catch all. It's the only way to do a netinstall for any of the desktops. The default package set is for a minimal install, a.k.a. Fedora Custom. The idea is to make it straightforward to get a minimum bootable system and then use dnf to build it up from there. Just for context, "Everything ISO" e.g. https://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-Rawhide-20220124.… This image defines "automatic" partitioning as btrfs. It's intrinsic to the image itself what the partitioning defaults to. It's not a function of the package set, i.e. choosing the Server package set doesn't get you Server's default partitioning. It's a little confusing, but it's working as designed, insofar as it has one. I cc'd Anaconda folks, because I'm not sure exactly where the "autopart" kickstart command gets defined these days. And also whether it's worth the effort to make autopart variable based on what edition or spin is selected in the Everything netinstaller? I'm not sure that it is. -- Chris Murphy

5 5

deprecate bootloader selection criterion
by Chris Murphy 24 Jan '22

24 Jan '22

https://fedoraproject.org/wiki/Fedora_35_Final_Release_Criteria#Bootloader_… "The installer must allow the user to choose which disk the system bootloader will be installed to, and to choose not to install one at all. " The "choose not to install one at all" doesn't really apply to UEFI. I wonder if we can just drop the whole criterion? (I'm still hopeful that one day the entirety of bootloader UI in the installer goes away.) So what you get is a system without an EFI system partition (or an existing ESP that isn't used), but all the bootloader stuff is put on the /boot volume. So it's not going to boot, which makes the installer's warnings true. The one thing *not* copied for some reason is the stub grub.cfg. But shim, grub, the real grub.cfg, and BLS snippets are all there - just on ext4 /boot which the firmware can't read so the system doesn't boot. It's harmless because the user signed up for this after all, but it's also kinda pointless. Reproduce steps: 1. Any that use Anaconda, from any release, but I just used this: Fedora-Workstation-Live-x86_64-Rawhide-20220119.n.0.iso 2. Boot it, launch installer 3. Installation destination keep all defaults 4. Click blue text at the bottom "Full disk summary and boot loader..." 5. Select the (in my case single) drive, click the "Do not install bootloader" button 6. Close and install -- Chris Murphy

5 6

the possibility to detect /home subvolume if fedora with default partition layout is detected.
by majid hussain 18 Jan '22

18 Jan '22

hi, firstly, thank you for a lovly installer. i'm blind and use orca the screen reader to aid me in installing fedora. too my question i'm a end user, could what was asked in the subject line be made a possibility? usecase, fedora install is completely inoperable and you wish to currently doo a repair install and you wish to retain your home subvolume. that is to say. 1. reinstall fedora lieveing your home subvolume alone and using that instead of formatting everything. this would lieve your userdata alone including your .files. you would need to make the same user account with the same name etc for this to work. how i see this in anaconda. when the installer is launched and you have gone through the language and keybord layout questions, when it getts to the partition stage and default /auto partitioning is selected a question could be asked doo you wish to keep your home partition/volume? I hope this makes sence. I would be greatful if you were to considder this. Majid Hussain

4 6

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel