Call for testing!

Call for testing!

[Ludovic Courtès]

[-- Attachment #1: Type: text/plain, Size: 836 bytes --]

Hi!

Please test the USB installation image at:

  http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz
  SHA1: 0d7f37d5b18f7b4437b447fc8a90b44c470f6e99
  http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz.sig

This is for x86_64 and the installation instruction still apply:

  https://www.gnu.org/software/guix/manual/html_node/System-Installation.html

… but tty2 in the installation image provides the latest version of the
manual, with slightly different examples (notably with an encrypted root
partition.)

I’m sorry that once again these arrive a day or two before the target
release date.  Please report back ASAP.

Reports of installations in VMs are interesting; bare-metal installation
are even better.

Thanks!  :-)

Ludo’.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: Call for testing!

[宋文武]

On 2015-11-03 22:17, ludo@gnu.org wrote:
> Hi!
> 
> Please test the USB installation image at:
> 
> 
> http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz
>   SHA1: 0d7f37d5b18f7b4437b447fc8a90b44c470f6e99
> 
> http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz.sig
> 
> This is for x86_64 and the installation instruction still apply:
> 
>   
> https://www.gnu.org/software/guix/manual/html_node/System-Installation.html
> 
> … but tty2 in the installation image provides the latest version of the
> manual, with slightly different examples (notably with an encrypted 
> root
> partition.)
Hi!  I test this install image using qemu by:

   qemu-system-x86-64 -enable-kvm -m 512 -hda guixsd-usb-install* -hdb 
disk.img

Follow the manual to setup encryted root, using the desktop.scm 
template, but
at the final step, it failed with:

   Path '/mnt/boot/grub' is not readable by GRUB on boot.
   Installation is impossible. Aborting.

   (can be reproduced by `grub-install /dev/sdb --boot-directory 
/mnt/boot')

After search, it seems that an un-encrypted boot partition is needed:

   
https://wiki.archlinux.org/index.php/Dm-crypt/Encrypting_an_entire_system#Preparing_the_boot_partition

So, I run fdisk to add a boot partiotion, finally install finished!

But it can't boot, I have to mount the encrypted root in Grub's cmdline:

   insmod cryptodisk
   insmod luks
   cryptomount hd0,msdos2

That's all I get so far.

Re: Call for testing!

[Ludovic Courtès]

宋文武 <iyzsong@openmailbox.org> skribis:

> Hi!  I test this install image using qemu by:
>
>   qemu-system-x86-64 -enable-kvm -m 512 -hda guixsd-usb-install* -hdb
> disk.img
>
> Follow the manual to setup encryted root, using the desktop.scm
> template, but
> at the final step, it failed with:
>
>   Path '/mnt/boot/grub' is not readable by GRUB on boot.
>   Installation is impossible. Aborting.
>
>   (can be reproduced by `grub-install /dev/sdb --boot-directory
> /mnt/boot')
>
> After search, it seems that an un-encrypted boot partition is needed:
>
>   https://wiki.archlinux.org/index.php/Dm-crypt/Encrypting_an_entire_system#Preparing_the_boot_partition
>
> So, I run fdisk to add a boot partiotion, finally install finished!
>
> But it can't boot, I have to mount the encrypted root in Grub's cmdline:
>
>   insmod cryptodisk
>   insmod luks
>   cryptomount hd0,msdos2

Bah, that got insufficient testing, so we’ll have to back up on this.  I
thought from http://bugs.gnu.org/19190 that we were done, but we’re not.

I’ve reverted recommendations of an encrypted root in the manual.  A bit
of a shame, but we’ll fix it soonish.

Thanks for the report!

Ludo’.

Re: Call for testing!

[Adam Pribyl]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2648 bytes --]

On Tue, 3 Nov 2015, Ludovic Courtès wrote:

> Hi!
>
> Please test the USB installation image at:
>
>  http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz
>  SHA1: 0d7f37d5b18f7b4437b447fc8a90b44c470f6e99
>  http://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.9.0pre.x86_64-linux.xz.sig
>
> This is for x86_64 and the installation instruction still apply:
>
>  https://www.gnu.org/software/guix/manual/html_node/System-Installation.html
>
> … but tty2 in the installation image provides the latest version of the
> manual, with slightly different examples (notably with an encrypted root
> partition.)
>
> I’m sorry that once again these arrive a day or two before the target
> release date.  Please report back ASAP.
>
> Reports of installations in VMs are interesting; bare-metal installation
> are even better.

I've tested on real hw with desktop config without crypto.

Notes:

1) doc: I would suggest to put in the doc (on tty2) to put the commands 
that should be used during installation on the separate lines

e.g.:
"
o install the system, you would:

     Configure the network, by running ifconfig eno1 up && dhclient eno1
"
should be
"
o install the system, you would:

     Configure the network, by running

       ifconfig eno1 up && dhclient eno1
"

This really helps when one is quickly scanning the installation procedure 
to find the desired commands.


2) I do sometimes see a guix deamon to spit a
"spurious SIGPOLL"
message. I do not know what it means.

3) I do see the process of installtion beeing from time to time aborted 
due to "can not resolve hydra.gnu.org" (sorry the message is not exact).

4) would be nice if installation image includes mouse (gpm) support - I 
did not found and easy way to copy the content from installation guide.

5) after default installation and reboot, grub loads the kernel, but the 
boot ends with
"waiting for "root" device to appear"
erroring out as it can not be found.
This was caused by the label "root" missing on /dev/sda1 partition, even 
thou it is specified in config.scm.

6) At first I had troubles with name resolution in the installed system. 
All address translation requests do end with "host not found". I have to 
explicitely ask with bind-utils host command, then the name is cached and 
I can use it. After some more package installation, that brought in the 
glibc or something and reboot it started to work OK?! (Related to 3?)

> Thanks!  :-)


I am very impressed with the GuixSD advancement. Thanks to all.


> Ludo’.

Adam Pribyl

Re: Call for testing!

[Ludovic Courtès]

Adam Pribyl <pribyl@lowlevel.cz> skribis:

> 1) doc: I would suggest to put in the doc (on tty2) to put the
> commands that should be used during installation on the separate lines
>
> e.g.:
> "
> o install the system, you would:
>
>     Configure the network, by running ifconfig eno1 up && dhclient eno1
> "
> should be
> "
> o install the system, you would:
>
>     Configure the network, by running
>
>       ifconfig eno1 up && dhclient eno1
> "
>
> This really helps when one is quickly scanning the installation
> procedure to find the desired commands.

Indeed.  Done in 235cba8.

> 2) I do sometimes see a guix deamon to spit a
> "spurious SIGPOLL"
> message. I do not know what it means.

It’s a harmless annoyance, but we should fix dmd to redirect the
daemon’s standard output.

> 3) I do see the process of installtion beeing from time to time
> aborted due to "can not resolve hydra.gnu.org" (sorry the message is
> not exact).

What networking driver were you using?  Normally, if dhclient works
correctly and there’s no network issue, that shouldn’t happen.

> 4) would be nice if installation image includes mouse (gpm) support -
> I did not found and easy way to copy the content from installation
> guide.

Good idea, done in ae7ffa9.  I set it to use the ‘ps2’ protocol, which
apparently works for USB HID mice.  I hope that’ll be fine.

> 5) after default installation and reboot, grub loads the kernel, but
> the boot ends with
> "waiting for "root" device to appear"
> erroring out as it can not be found.
> This was caused by the label "root" missing on /dev/sda1 partition,
> even thou it is specified in config.scm.

Yeah, the root partition must be created with ‘mkfs.ext4 -L the-label’.
I’ll make that command more clearly visible in “System Installation.”

> 6) At first I had troubles with name resolution in the installed
> system. All address translation requests do end with "host not
> found". I have to explicitely ask with bind-utils host command, then
> the name is cached and I can use it. After some more package
> installation, that brought in the glibc or something and reboot it
> started to work OK?! (Related to 3?)

This is weird.  Does the system use dhclient or wicd?  Is it WiFi or
wired?

Thanks a lot for your feedback.  It’s insightful, as usual!

Ludo’.

Call for testing!

[Ludovic Courtès]

[-- Attachment #1: Type: text/plain, Size: 507 bytes --]

Hello Guix!

Here’s an installation image for testing:

  http://www.fdn.fr/~lcourtes/software/guix/gsd-usb-install-20150127.x86_64-linux.xz
  http://www.fdn.fr/~lcourtes/software/guix/gsd-usb-install-20150127.x86_64-linux.xz.sig
  SHA1: db3cdaa6a4dbd4d84fc921f6526e237c55c7f9e6

It was built with:

  guix system disk-image --image-size=850MiB gnu/system/install.scm

From commit b1e95ec.

On #guix Joshua reported a successful installation with yesterday’s master.

Thanks,
Ludo’.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]