bug#37739: System install issue: 'You have a memory leak'

bug#37739: System install issue: 'You have a memory leak'

[Seswu M. Fafnan]

[-- Attachment #1: Type: text/plain, Size: 843 bytes --]

Hello.

I'm trying to install the full guix system on a pc but there's a problem.

When "populating '/mnt'", it bails with an error and claims:
'You have a memory leak (not released memory pool)'

A program named dtree is mentioned as well during what appears to be several attempts at having whatever it is succeed, then the install fails.

I can access other text terminals to read log files, but have not found any log in /var/log that resembles screen output, and did not find any guidance for this on the site/web either.

Output from inxi on system attached. I am not aware of any problems with the physical memory on this machine.

Install was 'normal' through usb stick, graphical install, english system, danish localization, full harddisk used, encrypted, with suggested layout.
There is no config.scm in /etc.

Kind regards,
  Uffe

[-- Attachment #2: inx.out --]
[-- Type: application/octet-stream, Size: 1336 bytes --]

System:
  Host: gnu Kernel: 5.1.2-gnu x86_64 bits: 64 Console: N/A Distro: unknown 
Machine:
  Type: Desktop Mobo: MEDION model: BTDD-LT v: 1.0
  UEFI: American Megatrends v: BTLTW08.106 date: 05/22/2014 
CPU:
  Topology: Quad Core model: Intel Celeron J1900 bits: 64 type: MCP 
  L2 cache: 1024 KiB 
  Speed: 1333 MHz min/max: 1333/2416 MHz Core speeds (MHz): 1: 1333 2: 1340 
  3: 2416 4: 2416 
Graphics:
  Device-1: Intel Atom Processor Z36xxx/Z37xxx Series Graphics & Display 
  driver: i915 v: kernel 
  Display: server: No display server data found. Headless machine? tty: N/A 
  Message: Advanced graphics data unavailable in console for root. 
Audio:
  Device-1: Intel Atom Processor Z36xxx/Z37xxx Series High Definition Audio 
  driver: snd_hda_intel 
  Sound Server: ALSA v: k5.1.2-gnu 
Network:
  Device-1: Realtek RTL810xE PCI Express Fast Ethernet driver: r8169 
  IF: enp2s0 state: up speed: 100 Mbps duplex: full
Drives:
  Local Storage: total: 938.97 GiB used: 5.37 GiB (0.6%) 
  ID-1: /dev/sda vendor: Toshiba model: DT01ACA100 size: 931.51 GiB 
  ID-2: /dev/sdb type: USB vendor: Intenso model: Alu Line size: 7.46 GiB 
Sensors:
  System Temperatures: cpu: 40.0 C mobo: N/A 
  Fan Speeds (RPM): N/A 
Info:
  Processes: 106 Uptime: 1h 40m Memory: 3.73 GiB used: 601.2 MiB (15.7%) 
  Init: N/A Shell: bash inxi: 3.0.33 

bug#37739: System install issue: 'You have a memory leak'

[Ludovic Courtès]

Hi Seswu,

"Seswu M. Fafnan" <hadaqada@gmail.com> skribis:

> I'm trying to install the full guix system on a pc but there's a problem.
>
> When "populating '/mnt'", it bails with an error and claims:
> 'You have a memory leak (not released memory pool)'

I’ve never seen this message before, but looking at the intertubes, it
may be a bug in GRUB related to encryption:

  https://github.com/NixOS/nixpkgs/issues/21830

Could you explain how to reproduce the issue?  In particular, what
partitioning scheme and encryption setup did you choose?

Thanks in advance,
Ludo’.

bug#37739: System install issue: 'You have a memory leak'

[Seswu M. Fafnan]

Hello Ludovic :)

Thank you for responding.
Late in the evening, with help from leoprikler at irc #guix, a workaround was found; going manual, which had until then been too overwhelming for me to face.
I just wrote a bit on that as extra information to the bug report, but will quote it below also.

Reproduction is to use the graphical installer; from the beginning:
English, Denmark, Graphical installer;
Europe, Copenhagen, Danish, Danish (no dead keys)
Guided - entire disk with encryption
ATA scsi (choosing my harddisk)
Everything is one partition

Resulting partitioning scheme looked like:
/dev/sda
	/boot/efi
		576MB fat32
		boot,esp
	/
		1000GB ext4
		cryptroot

I did not encounter the same problem doing the population manually.
The output from the manual command and from the graphical install is not the same, so I can't really compare what happened blow-by-blow.
I tried the graphical approach thrice, getting the issue, then following advice did the manual command - actually in an attempt to secure the screen output - and it worked the first time I did it.

I hope that this helped in determining the source of the problem :]

Kind regards,
  Uffe

---------------------------------------
With the help of #guix, a workaround was found;
A manual continuation of the install succeeded where the graphical install did not.
'guix system init /mnt/etc/config.scm /mnt' resulted in successful population of /mnt, and in what seems to me to be a functioning system.
---------------------------------------

On Sun, 13 Oct 2019 22:47:53 +0200
Ludovic Courtès <ludo@gnu.org> wrote:

> Hi Seswu,
> 
> "Seswu M. Fafnan" <hadaqada@gmail.com> skribis:
> 
> > I'm trying to install the full guix system on a pc but there's a problem.
> >
> > When "populating '/mnt'", it bails with an error and claims:
> > 'You have a memory leak (not released memory pool)'
> 
> I’ve never seen this message before, but looking at the intertubes, it
> may be a bug in GRUB related to encryption:
> 
>   https://github.com/NixOS/nixpkgs/issues/21830
> 
> Could you explain how to reproduce the issue?  In particular, what
> partitioning scheme and encryption setup did you choose?
> 
> Thanks in advance,
> Ludo’.


-- 
Seswu M. Fafnan <hadaqada@gmail.com>

bug#37739: System install issue: 'You have a memory leak'

[Ludovic Courtès]

Hi Seswu,

"Seswu M. Fafnan" <hadaqada@gmail.com> skribis:

> Late in the evening, with help from leoprikler at irc #guix, a workaround was found; going manual, which had until then been too overwhelming for me to face.
> I just wrote a bit on that as extra information to the bug report, but will quote it below also.
>
> Reproduction is to use the graphical installer; from the beginning:
> English, Denmark, Graphical installer;
> Europe, Copenhagen, Danish, Danish (no dead keys)
> Guided - entire disk with encryption
> ATA scsi (choosing my harddisk)
> Everything is one partition
>
> Resulting partitioning scheme looked like:
> /dev/sda
> 	/boot/efi
> 		576MB fat32
> 		boot,esp
> 	/
> 		1000GB ext4
> 		cryptroot
>
> I did not encounter the same problem doing the population manually.
> The output from the manual command and from the graphical install is not the same, so I can't really compare what happened blow-by-blow.
> I tried the graphical approach thrice, getting the issue, then following advice did the manual command - actually in an attempt to secure the screen output - and it worked the first time I did it.

So doing a similar installation using the “manual” installation process
worked, right?

Did you end up with an OS config file different from the one produced by
the graphical installer?  Could you post that config file that
eventually worked so we can compare it to what the installer produces?

Thanks for the update!

Ludo’.

bug#37739: System install issue: 'You have a memory leak'

[Seswu M. Fafnan]

[-- Attachment #1: Type: text/plain, Size: 2947 bytes --]

Hello Ludovic,

Yes, I did did a similar install - in fact I was simply 'continuing' the install manually from the point where the graphical install failed.

So, I was using the config file that the graphic config had produced.
I still am.
There is some complaints from the package manager about localization, and also something about putting my .guix-profile directory into an environment variable. I've tried to figure out what the latter is about, but I don't know how it is -supposed- to work, so I'm not sure what to do about it. From what I hear on irc, the profile and probably everything in the .guix-profile dir is repeatedly generated, which means it's not -there- I should correct things probably. So I'm a bit confused on that point.
I also have some trouble with the window manager I chose, i3. It seems it didn't fully install. It's called up, and works.. the status bar didn't though, not until I installed it manually, and I still haven't gotten dmenu to work properly.

Regardless.. those three things are probably besides the point/bug. The system got up and started, it works, I can log into it, and if I knew what I was doing I could probably remedy the remaining things in a snap.

I've attached the config file I used. (generated by the graphical installer)

Uffe


 On Wed, 16 Oct 2019 22:30:57 +0200
Ludovic Courtès <ludo@gnu.org> wrote:

> Hi Seswu,
> 
> "Seswu M. Fafnan" <hadaqada@gmail.com> skribis:
> 
> > Late in the evening, with help from leoprikler at irc #guix, a workaround was found; going manual, which had until then been too overwhelming for me to face.
> > I just wrote a bit on that as extra information to the bug report, but will quote it below also.
> >
> > Reproduction is to use the graphical installer; from the beginning:
> > English, Denmark, Graphical installer;
> > Europe, Copenhagen, Danish, Danish (no dead keys)
> > Guided - entire disk with encryption
> > ATA scsi (choosing my harddisk)
> > Everything is one partition
> >
> > Resulting partitioning scheme looked like:
> > /dev/sda
> > 	/boot/efi
> > 		576MB fat32
> > 		boot,esp
> > 	/
> > 		1000GB ext4
> > 		cryptroot
> >
> > I did not encounter the same problem doing the population manually.
> > The output from the manual command and from the graphical install is not the same, so I can't really compare what happened blow-by-blow.
> > I tried the graphical approach thrice, getting the issue, then following advice did the manual command - actually in an attempt to secure the screen output - and it worked the first time I did it.
> 
> So doing a similar installation using the “manual” installation process
> worked, right?
> 
> Did you end up with an OS config file different from the one produced by
> the graphical installer?  Could you post that config file that
> eventually worked so we can compare it to what the installer produces?
> 
> Thanks for the update!
> 
> Ludo’.


-- 
Seswu M. Fafnan <hadaqada@gmail.com>

[-- Attachment #2: config.scm --]
[-- Type: application/octet-stream, Size: 1747 bytes --]

;; This is an operating system configuration generated
;; by the graphical installer.

(use-modules (gnu))
(use-service-modules desktop networking ssh xorg)

(operating-system
  (locale "en_DK.utf8")
  (timezone "Europe/Copenhagen")
  (keyboard-layout
    (keyboard-layout "dk" "nodeadkeys"))
  (bootloader
    (bootloader-configuration
      (bootloader grub-efi-bootloader)
      (target "/boot/efi")
      (keyboard-layout keyboard-layout)))
  (mapped-devices
    (list (mapped-device
            (source
              (uuid "d1a5a5b6-7205-4005-9f55-989d33f3264e"))
            (target "cryptroot")
            (type luks-device-mapping))))
  (file-systems
    (cons* (file-system
             (mount-point "/boot/efi")
             (device (uuid "35CD-F935" 'fat32))
             (type "vfat"))
           (file-system
             (mount-point "/")
             (device "/dev/mapper/cryptroot")
             (type "ext4")
             (dependencies mapped-devices))
           %base-file-systems))
  (host-name "tabula-rasa")
  (users (cons* (user-account
                  (name "seswu")
                  (comment "Seswu")
                  (group "users")
                  (home-directory "/home/seswu")
                  (supplementary-groups
                    '("wheel" "netdev" "audio" "video")))
                %base-user-accounts))
  (packages
    (append
      (list (specification->package "i3-wm")
            (specification->package "nss-certs"))
      %base-packages))
  (services
    (append
      (list (service openssh-service-type)
            (service tor-service-type)
            (set-xorg-configuration
              (xorg-configuration
                (keyboard-layout keyboard-layout))))
      %desktop-services)))

bug#37739: Acknowledgement (System install issue: 'You have a memory leak')

[Seswu M. Fafnan]

With the help of #guix, a workaround was found;
A manual continuation of the install succeeded where the graphical install did not.
'guix system init /mnt/etc/config.scm /mnt' resulted in successful population of /mnt, and in what seems to me to be a functioning system.





On Sun, 13 Oct 2019 19:06:02 +0000
help-debbugs@gnu.org (GNU bug Tracking System) wrote:

> 37739@debbugs.gnu.org

-- 
Seswu M. Fafnan <hadaqada@gmail.com>