Manjaro Architect Part ???: XFCE Edition.

With lvm+luks, password was asked twice. It might be triggered by the disk filter error or there is something wrong with the installer.

Now testing if adding --iter-time 500 affects the grub boot time. EDIT: using a lower iter-time made no difference in decryption time, it stayed stably in 21s

Very nice! Thank you.

Allow me to run the same tests tonight before making a decision, I was convinced --iter-time played a role in all of this. A quick "--itter-time slow boot" Google search reveals a pattern.

Feels nice to know it's not imaginary hahaha.

I think I know why grub is not installed on the first go. I'll just have to adjust the order of operations in the grub installation, and it should then work with one go.

Sure, my googlefu suggests the same. If you can find where the flag should be applied, it would be awesome.

Okay, I think I got the disk filter error and need to install grub twice solved. Testing before uploading changes.

Next issue is why keyfile was not working with lvm root. I'll make a new release when that is fixed. After that I'll see if I can make the decryption faster.

Give me a chance to work this bug. I'd appreacitate it, it's been hustling my patience for awhile.

Okay, issues solved.

The reason lvm+luks asks password twice is that I reused an old luks container. M-a won't add another key if cyptsetup -luksDump shows more than 1 key already. This is a feature to avoid creating new keys when reinstalling the bootloader. I think I need to reconsider the feature.

EDIT. Fixes pushed to all branches

1 Like
  • GRUB still needs to be installed twice.
  • After installing KDE Minimal, making some minor changes and rebooting, nothing would launch.
    • Changes are: Scaled desktop and connecting to WIFI.
    • Scaling is broken (disparity between Application launcher and panel).
  • ■■■■ my life! Introduced the --iter-time parameter and now GRUB takes 2 minutes to unlock.
  • fg: current: no such job. M-A's Chroot cannot be escaped.

Proprietary drivers (NVIDIA MX150) and LVM were involved.

Did you first quit the chroot shell with exit or ctrl+d?

exit locks-up M-A; entering fg afterwards is ineffective.

CTRL + d isn't mentioned, so I did not bother.

Well, it does the same as exit anyway. Thank you for reporting.

Btw, did you use 0.9.25?

Yes.

The most serious offender being unresponsive program launches. Probably tied with GRUB slow unlock, though it's just a matter of figuring out what makes it tick.

Frankly I don't think it's fair for you to be tasked with addressing all of these problems by yourself. ISO-profiles' issue was not yours to carry, yet you fixed it anyways.

I come bearing gifts, thanks to the program that could --Yakuake.

Here are the errors!
Launching palemoon:

(pale moon:3199): Gtk-WARNING **: 00:18:47.904 Locale not supported by C library.
                Using the fallback 'C' locale.
No protocol specified
Error: cannot open display: :0

Launching konsole:

No protocol specified
qt.qpa.xcb: could not connect to display :0 
qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "" even though it was found.
This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.

Available platform plugins are: wayland-org.kde.kwin.qpa, eglfs, linuxfb, minimal, minimalegl, offscreen, vnc, xcb.

`konsole` terminated by signal SIGABRT (Abort)

Launching Falkon = same Qt error above.

Launching manjaro-hello:

Warning: Application Browser plugin not found : No module named 'application_utility'
No protocol specified
Unable to init server: Could not connect: Connection refused

(manjaro-hello:3472): Gtk-WARNING **: 00:39:09.364: Locale not supported by C library.
                Using the fallback 'C' locale.
No protocol specified
Unable to init server: Could not connect: Connection refused 
Traceback (most recent call last):
  File "/usr/bin/manjaro-hello", line 347, in <module> 
    hello = hello()
  File "/usr/bin/manjaro-hello", line 31, in __init__
    Gtk.Window.__init__(self, title=" Manjaro Hello", border_width=6)
  File "/usr/lib/python3.7/site-packages/gi/overrides/Gtk.py", line 519, in __init__
    "Gtk couldn't be initialized. "
RuntimeError: Gtk couldn't be initialized. Use Gtk.init_check() if you want to handle this case.

Parts of the session are in español!

Did you remember to generate locales when installing? Some programs fail to launch if locales are not generated.

Also, the error indicates that $DISPLAY is not set. Maybe an issue with sddm?

Unfortunately I'm not that familiar with qt, maybe there should be a separate thread about this issue?

Yeah, I'll spin one by noon.

Hmm, exiting tty2 returns a message on a black backdrop in bold, stating " The screen locker is broken and unlocking is not possible anymore".

Found a thread in SUSE Forums with promissing solutions to both problems!

/tmp is full, and if that's the case finding out why it's now a problem.

@Chrysostomus

--iter-time is ignored if entered in the "Define Key-Size and Cypher" field, correct?

That's where I put it, (i.e. --iter-time 10000), which somehow generated 10 million iterations for key-slot 0 (User-generated passphrase) and 1 million iterations for key-slot 1 (Automatically generated, introduced post-GRUB). Perhaps I slipped an extra 0, not totally sure.

--pbkdf-force-iterations is most effective in limiting number of iterations as it introduces a hard-cap, meeting desired value in time. --iter-time defines number of ms spent processing PBKDF, according to The Bible. The former reduced GRUB unlock to 1 second. Not very secure in the grand scheme of things, but it can be done.

Found a good piece on the effects of --pbkdf-force-iterations parameter.

1 Like

Thanks! That will definitely be useful. Maybe I can add it as an alternative in the luks menu.

Let me see if I can get --iter-time to behave; inserting it in Define Key-Size and Cypher field produces negative results. Why is SHA256 default instead of SHA512?

Who maintains XFCE/KDE Editions?

I think @philm maintains xfce. Not sure who maintains kde.

I now added an option for quick and insecure LUKS in manjaro-architect, dubbed Express LUKS. It seems to provide quick fde boot.

Ehh, it's alright. Restored images of working installs, if issue persists after next stable release I might dig deeper.

Boss! Wish there was additional documentation on these two, seemingly identical flags. --pbkdf-force-iterations is not a hardware ambiguous resolve (What I thought --iter-time could acomplish). Forcefully effective regardless, will spend some more time reading. :slightly_smiling_face:

Development version dev5

1 Like