4.0.1-RC2 Boot loop after install

[John Goold]

Attached is screenshot, taken under my current OS, showing OS and hardware info.

After spending much too much time trying to track the problem down (using the 4.0, 4.0.1-RC1 and 4.0.1-RC2 ISOs) I discovered why getting the installer to run was failing...

I had to unplug my external monitor (connected via an HDMI port).

I was then able to boot the install DVD and install to an external USB (SSD) drive (Seagate 2 TB). The install completed (supposedly successfully), but attempts to boot from the USB drive fail.

The boot process starts, with text being displayed starting in the top left corner of the screen. It progresses to a point, then the screen goes black and my computer starts to reboot.

I have searched the mailing list and have failed to find a solution (hours spent doing this). A lot of people seem to end up in boot-loops, using various hardware.

The attached file shows the hardware. The following information about the BIOS/Firmware may be relevant:

* Legacy Boot is enabled
* Virtualization Technology is enabled

During the install I setup a user account. I did not enable disk encryption (I will leave that until after I can get Qubes to boot).

Comment: This boot-loop problem (or similar boot-loop problems) seems to be a major issue with installing Qubes 4.x. Each time I come across a posting about it, there seem to be different suggestions (some of which work on the particular hardware involved) and some of which do not.

I believe that I tried R3.1 about a year or so ago and that it booted alright. I cannot remember why I did not follow through on adopting Qubes (if I could not get my external monitor working, that would be a deal-breaker).

Suggestions would be appreciated. I will provide any additional information I am capable of.

[John Goold]

I should have been clearer in indicating how far the boot process got:

The text messages were displayed starting in the top left corner of the screen until the screen cleared and the Qubes Q-logo was displayed with a progress bar underneath.

The boot process continued until the progress indicator was about 1/4 to 1/3 of the way across, then the screen went black and my computer starts to reboot.

[awokd]

John Goold:

> I should have been clearer in indicating how far the boot process got:
>
> The text messages were displayed starting in the top left corner of the screen until the screen cleared and the Qubes Q-logo was displayed with a progress bar underneath.
>
> The boot process continued until the progress indicator was about 1/4 to 1/3 of the way across, then the screen went black and my computer starts to reboot.
>

Did the 4.0 installer reboot in the same spot? I've heard of this
happening with newer kernels sometimes, but the one in the original
release is older. Also, if you're getting that far there might be
something in the log files. Boot in rescue mode if you can (or another
distribution), mount the system disk and check.

[John Goold]

@awokd Thank you very much for responding — I do appreciate the effort.

Here is my attempt to find the log file(s). I am booted in another distribution (Linux Mint 19.1 64-bit Cinnamon):

I used "sudo" to become root and then:

root@JRGsHPSpectre:/dev/qubes_dom0# lvmdiskscan
/dev/qubes_dom0/swap [ 7.48 GiB]
/dev/sda1 [ 200.00 MiB]
/dev/sda2 [ <27.95 GiB]
/dev/sda3 [ 428.30 GiB]
/dev/sda4 [ 9.31 GiB]
/dev/sdb1 [ 500.00 MiB]
/dev/sdb2 [ <1.82 TiB] LVM physical volume
1 disk
5 partitions
0 LVM physical volume whole disks
1 LVM physical volume
root@JRGsHPSpectre:/dev/qubes_dom0#

So there is the Qubes swap partition (7.48 GiB). /dev/sdb2 is the USB disk drive that was the target of the Qubes install.

lvdisplay >> ~john/temp/LVM-Info.txt # result below

--- Logical volume ---
LV Name pool00
VG Name qubes_dom0
LV UUID fWWjqo-5OdP-J5pO-obIu-Is0u-KekI-w5Dj26
LV Write Access read/write
LV Creation host, time localhost.localdomain, 2018-12-20 13:49:26 -0330
LV Pool metadata pool00_tmeta
LV Pool data pool00_tdata
LV Status NOT available
LV Size <1.80 TiB
Current LE 470795
Segments 1
Allocation inherit
Read ahead sectors auto

--- Logical volume ---
LV Path /dev/qubes_dom0/root
LV Name root
VG Name qubes_dom0
LV UUID EBmNyt-sCQi-5G4V-EpQQ-GLuJ-SCCY-MnWssJ
LV Write Access read/write
LV Creation host, time localhost.localdomain, 2018-12-20 13:49:27 -0330
LV Pool name pool00
LV Status NOT available
LV Size <1.80 TiB
Current LE 470795
Segments 1
Allocation inherit
Read ahead sectors auto

--- Logical volume ---
LV Path /dev/qubes_dom0/swap
LV Name swap
VG Name qubes_dom0
LV UUID aLPIfF-1EHL-5a1X-ejaj-DNdK-yHk2-TCLZXH
LV Write Access read/write
LV Creation host, time localhost.localdomain, 2018-12-20 13:49:28 -0330
LV Status available
# open 0
LV Size 7.48 GiB
Current LE 1915
Segments 1
Allocation inherit
Read ahead sectors auto
- currently set to 256
Block device 253:0

This appears to show 3 logical volumes: pool00, root and swap (which appears to be mounted). Note that pool00 and root both have an "LV Status" of "NOT available".

I created a mount point, /mnt/lvroot and then attempted to mount the root logical volume:

root@JRGsHPSpectre:/dev/qubes_dom0# ls -l /mnt
total 4
drwxr-xr-x 2 root root 4096 Dec 22 17:23 lvroot
root@JRGsHPSpectre:/dev/qubes_dom0#

root@JRGsHPSpectre:/dev/qubes_dom0# mount /dev/qubes_dom0/root /mnt/lvroot -o ro,user
mount: /mnt/lvroot: special device /dev/qubes_dom0/root does not exist.
root@JRGsHPSpectre:/dev/qubes_dom0#

Well, so much for that :(

Let us see if we can make the other logical volumes active:

root@JRGsHPSpectre:/dev/qubes_dom0# vgchange --activate y --force qubes_dom0
/usr/sbin/thin_check: execvp failed: No such file or directory
Check of pool qubes_dom0/pool00 failed (status:2). Manual repair required!
/usr/sbin/thin_check: execvp failed: No such file or directory
1 logical volume(s) in volume group "qubes_dom0" now active
root@JRGsHPSpectre:/dev/qubes_dom0#

Well, I guess I am out of my depth with LVM or the install did not create the LVM group/volumes/<whatever> correctly. :(

[awokd]

John Goold:

> Well, I guess I am out of my depth with LVM or the install did not create the LVM group/volumes/<whatever> correctly. :(
>

Those LVM commands don't look quite right. Try the ones mentioned here:
https://unix.stackexchange.com/questions/232905/lvm-mount-rescue-mode
for example. If you do get root to mount, look at /var/log/boot.log in
there, and maybe /var/log/xen/console/hypervisor.log.

[John Goold]

I tried those, they only repeat the same results I was already getting.

What bothers me is that all the discussions I have found so far about LVM discuss physical drives (PV - Physical Volume), Volume Groups (VG) and Logical Volumes (LV). An example is https://www.digitalocean.com/community/tutorials/an-introduction-to-lvm-concepts-terminology-and-operations

However, they do not discuss "Pools". My output from the lvs command has a column I do not see in any of the discussions labelled "Pool":

root@JRGsHPSpectre:~# lvs
LV VG Attr LSize Pool
pool00 qubes_dom0 twi---tz-- <1.80t
root qubes_dom0 Vwi---tz-- <1.80t pool00
swap qubes_dom0 -wi-a----- 7.48g
root@JRGsHPSpectre:~#

I deleted the remaining column labels as there was nothing listed under them. Notice that the two Logical Volumes that are not mounted automatically (and which I am having problems with) each have an attribute that "swap" (the LV that is mounted automatically) does not have: "pool00" has the "t" attribute and "root" has the "V" attribute. Also, those two LVs have exactly the same size.

I am guessing that "pool00" is some higher level of management and that "root" is allocated in "pool00" — but it is purely a guess.

So far I have not encountered information about "pools" with regard to LVM. It looks like I need to understand these in order to mount the "root" LV (or at least find the necessary commands to work with pools).

[John Goold]

=======================================

I may have found something, "thin Provisioned Volumes": https://www.linuxtechi.com/thin-provisioned-logical-volumes-centos-7-rhel-7/

I have only skimmed the article. I need to go through it carefully.

[John Goold]

This thread is getting verbose, so I have replied to the original post and will attempt a brief summary of the rest of the thread (for context):

Determining what is happening would be facilitated by seeing any entries in log files (assuming the boot got far enough to log anything).

That means checking files on the USB drive used as the target of the install and which causes the boot-loop when attempting to boot.

Since the boot is failing, I cannot look at the log files under the booted Qubes OS, so instead I attempted to look for the log files when booted into another OS (Linux Mint 19.1).

Qubes is using LVM to handle allocating disk space (presumably to facilitate being able to add additional physical disks to an existing Qubes install). There appeared, at first glance to be 3 Logical volumes:

pool00
root
swap

Linux Mint mounted the LV "swap" automatically, but not the other two. The other two appear not to be "activated" and mount attempts failed. Attempts to "activate" the LVs fail.

After searching the Net for information on LVM, I came across an article that helped me understand the Qubes setup better…

There is one Logical Volume Group called "qubes_dom0".
Within that there is a Logical Volume, "swap", that is detected and mounted automatically by my Linux Mint installation.
Additionally, there is a "Thin Pool" allocated that uses up the rest of the space in the Volume Group. It is distinguished by information displayed by the lvdisplay command ("LV Pool metadata" and "LV Pool data").

Within that "thin pool", a logical volume, "root" has been created that uses all the disk space currently assigned.

Unfortunately, that knowledge has not helped me mount the LV, "root". The mount fails. It is still not clear whether the Qubes install has done (or failed to do) something to the LVM setup or whether I have just failed to understand how to activate the thin pool and the contained logical volume.

[Marek Marczykowski-Górecki]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Yes, that's right.

- From what I've seen in this thread, you did it right, but the system you
used didn't support thin volumes. You can try Qubes installation image,
there is recovery mode ("Rescue" in boot menu in legacy mode).


Other things you can try is to press ESC during boot to see more
messages than just progress bar. If that doesn't really help, try
editing boot entry in grub and remove "quiet" and "rhgb" options from
there. This should give you more details when exactly system reboots.

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEhrpukzGPukRmQqkK24/THMrX1ywFAlwlLecACgkQ24/THMrX
1yw98wf/e5BWkvOJQjrDpiB4DDpmRKAWvmZ3b/NssWOZgtBTMdYZxcRHaguoSQx2
hrMUKr5fLT3xy5fQYzt/OSjWjEsfbYoLfG8hx32+Zp0WDTpfSjon85/HerYlb8TW
pHS0lxT4y0sTDPx8HHy4O+0tFiYnHGeG9+JEpRx9JTxxJc6jX5hiw/DlaIldO9AD
/Qdi3aQjpU7qFMeQq0MebhHcaPME57lc87SkFlHhKnjEC8CsVpYSRiEKRx4ufREn
7GKeiayK0bEqm9HZWsVJyn4XBTtEiL9kE5VK6RY6pv5Xx7cOEqzQiMchpSIa5/rQ
jcM+S2IXxzL+4t2epl8//xMziDFaGQ==
=s/Kn
-----END PGP SIGNATURE-----

[John Goold]

The messages scroll past much too quickly for me and disappear almost immediately the scrolling stops (i.e. as soon as the reboot starts).

The last message (the only one I was able to read) said: "Starting switch boot" (it may have been "Start …" or "Started …".

I do not know if that helps or if it is simply the message that says, "… OK, may as well reboot".

I did wonder if the proble was because I was using a 2TB USB drive (SSD), so I also tried using a 500 GB external USB drive (HDD) — same result.

My wife bought me a nice new computer for Christmas — it is an Intel NUC7i7. I'm just waiting for my local computer store to get in some more 16 GB DIMMs before trying it out (I am putting in the maximum of 32GB (I have the 500 GB SSD installed already) I'll see if I get any different results with a different computer.

[John Goold]

The boot-loop problem appears to be a problem with the HP Spectre X360.

I have the NUC7i7BNH up and running. I was able to install R1.0.1 (stable) on it. Since it will become my primary computer, I no longer care about the HP Laptop (at least in terms of installing Qubes on it!). At present I am reliant on the HP during the transition to Qubes — it is a bit of a bumpy road.