Bug 2168506

Let's look at this for RHEL 9.3.  The current option is
not optimal but perfectly usable.

And remember UUIDs -- you can use UUIDs with "--key", and (as UUID says in the name) those should uniquely identify the device.

from: /dev/mapper/rhel_bootp--73--75--123-root
to: /dev/rhel_bootp-73-75-123/root

Rich wrote:

'In general /dev/VG/LV is the same as /dev/mapper/VG-LV.  If either "VG" or "LV" contain single "-" characters, they get doubled ("--").'

I thought we'd need a regex with capturing parens for translating the /dev/mapper/VG-LV format to /dev/VG/LV, but it seems that collapsing "--" to "-", plus ensuring there is precisely one standalone "-" before the collapsing, should suffice. AFAICT this should be possible to do with a single loop pass over VG-LV, with some tracking pointers maintained.

[libguestfs-common PATCH 0/2] recognize "/dev/mapper/VG-LV" with "--key"
Message-Id: <20230515174924.290409-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-May/031497.html

[v2v PATCH 0/2] test "/dev/mapper/VG-LV" with "--key"
Message-Id: <20230515175529.290724-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-May/031501.html

[libguestfs-common PATCH v2 0/2] recognize "/dev/mapper/VG-LV" with "--key"
Message-Id: <20230518130942.288152-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-May/031571.html

(In reply to Laszlo Ersek from comment #7)
> [libguestfs-common PATCH v2 0/2] recognize "/dev/mapper/VG-LV" with "--key"
> Message-Id: <20230518130942.288152-1-lersek>
> https://listman.redhat.com/archives/libguestfs/2023-May/031571.html

Commit range 38e6988c1864..b636c3f20a1b.

(In reply to Laszlo Ersek from comment #6)
> [v2v PATCH 0/2] test "/dev/mapper/VG-LV" with "--key"
> Message-Id: <20230515175529.290724-1-lersek>
> https://listman.redhat.com/archives/libguestfs/2023-May/031501.html

Commit range e83de8abe6c5..3060af01e87f.

[libguestfs PATCH 0/3] test "/dev/mapper/VG-LV" with "--key"
Message-Id: <20230519140849.310774-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-May/031584.html

[guestfs-tools PATCH 0/3] test "/dev/mapper/VG-LV" with "--key"
Message-Id: <20230519155507.369494-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-May/031589.html

(In reply to Laszlo Ersek from comment #10)
> [libguestfs PATCH 0/3] test "/dev/mapper/VG-LV" with "--key"
> Message-Id: <20230519140849.310774-1-lersek>
> https://listman.redhat.com/archives/libguestfs/2023-May/031584.html

Commit range 692802bf96e0..32408a9c3616.

(In reply to Laszlo Ersek from comment #11)
> [guestfs-tools PATCH 0/3] test "/dev/mapper/VG-LV" with "--key"
> Message-Id: <20230519155507.369494-1-lersek>
> https://listman.redhat.com/archives/libguestfs/2023-May/031589.html

Commit range 67647b883e13..569bd1dd29da.

Rich, for backporting purposes, should I clone this BZ for libguestfs (guestfish -i) and guestfs-tools (virt-inspector) as well? Thanks!

Whoops, setting needinfo as well now :)

Yes please, we will need separate bugs for every RHEL component that has to be changed.

[v2v PATCH] test-data/phony-guests: fix prerequisite list of "fedora-luks-on-lvm.img"
Message-Id: <20230619162729.153334-1-lersek>
https://listman.redhat.com/archives/libguestfs/2023-June/031856.html

(In reply to Laszlo Ersek from comment #18)
> [v2v PATCH] test-data/phony-guests: fix prerequisite list of "fedora-luks-on-lvm.img"
> Message-Id: <20230619162729.153334-1-lersek>
> https://listman.redhat.com/archives/libguestfs/2023-June/031856.html

Commit 13a6f4b9686e.

Development Management has reviewed and declined this request. You may appeal this decision by reopening this request.

I've pushed the following commits to the upstream rhel-9.3 branch (all cherry-picks from the master branch):

     1  6dea82d823c3 Update common submodule
     2  1d69132b7b72 update common submodule
     3  2558084d081c LUKS-on-LVM conversion test: rename VGs and LVs
     4  c8902c551014 LUKS-on-LVM conversion test: test /dev/mapper/VG-LV translation
     5  10192f8ee3a7 test-data/phony-guests: fix prerequisite list of "fedora-luks-on-lvm.img"

This bug is not stale. It is blocked, just on something that's not another RHBZ (at this point).

I'd be surprised if resubmitting the build works.  It seems something is going
very wrong when patching the paravirt instructions in the kernel.  Kernel or TCG bug?

(In reply to Richard W.M. Jones from comment #33)
> I'd be surprised if resubmitting the build works.  It seems something
> is going very wrong when patching the paravirt instructions in the
> kernel.  Kernel or TCG bug?

Yes, either guest kernel or TCG bug (or both!)

I expect there's a fair chance for the resubmitted build to succeed, for
the following reason:

- In c9s Koji task 2378814 (comment#24) and Brew task 53460798
  (comment#25), the same QEMU version was used (17:8.0.0-5.el9), and the
  same guest kernel version was used (5.14.0-330.el9.x86_64).

- In the c9s Koji task, the appliance kernel ran 6 times successfully.
  The first two runs are for image preparations (windows.img,
  fedora.img), the third run is a virt-v2v null conversion open-coded in
  the SPEC file, the fourth run is another image preparation
  (fedora-luks-on-lvm.img). The fifth and sixth runs are not logged as
  verbosely; they are from "test-v2v-fedora-luks-on-lvm-conversion.sh",
  from the test suite -- those correspond to two virt-v2v null
  conversions.

- In the Brew task, the appliance kernel ran 2 times successfully
  (windows.img, fedora.img) , it is the virt-v2v null conversion that's
  open-coded in the SPEC file where the appliance kernel crashes.

So it does not look deterministic. We can file a new bug later if
needed, but for now I want to complete the backport here.

Ah, the friendly bots have filed a new RHBZ about the build failure for us:

https://bugzilla.redhat.com/show_bug.cgi?id=2216425

That allows for some better dependency tracking.

Definitely a QEMU (TCG) or guest kernel bug: in the latest Brew build (comment 36), the first *four* appliance kernel boots succeeded, but then one of the last two (= 5th / 6th) ones hung -- the Brew build has been running for ~50 minutes now, and the log is stuck in "test-v2v-fedora-luks-on-lvm-conversion.sh".

(In reply to Laszlo Ersek from comment #38)
> Definitely a QEMU (TCG) or guest kernel bug: in the latest Brew build
> (comment 36), the first *four* appliance kernel boots succeeded, but then
> one of the last two (= 5th / 6th) ones hung -- the Brew build has been
> running for ~50 minutes now, and the log is stuck in
> "test-v2v-fedora-luks-on-lvm-conversion.sh".

--> https://bugzilla.redhat.com/show_bug.cgi?id=2216496

I had a loop running this all day yesterday, on RHEL 9, with all the same
packages installed, and it didn't fail at all, not that this proves anything much:

while LIBGUESTFS_BACKEND_SETTINGS=force_tcg libguestfs-test-tool -t 60 >& /tmp/1 ; do echo -n . ; done

Verify the bug with below builds:
virt-v2v-2.3.4-3.el9.x86_64
libguestfs-1.50.1-6.el9.x86_64
libvirt-libs-9.4.0-1.el9.x86_64
qemu-img-8.0.0-5.el9.x86_64
nbdkit-server-1.34.1-1.el9.x86_64
libnbd-1.16.0-1.el9.x86_64


Steps:
1.Prepare a guest with luks encrypted lvm on VMware
# lsblk
NAME                                            MAJ:MIN RM  SIZE RO TYPE  MOUNTPOINTS
sda                                               8:0    0   16G  0 disk  
├─sda1                                            8:1    0    1G  0 part  /boot
└─sda2                                            8:2    0   15G  0 part  
  ├─rhel_bootp--73--75--123-root                253:0    0 13.4G  0 lvm   
  │ └─luks-cf19e4ac-3f54-450b-b801-72e539a8f870 253:2    0 13.4G  0 crypt /
  └─rhel_bootp--73--75--123-swap                253:1    0  1.6G  0 lvm   [SWAP]
sr0     
                     
2.Convert the guest from VMware by v2v
#  virt-v2v -ic vpx://administrator%40vsphere.local.213.93/data/10.73.212.38/?no_verify=1 -it vddk -io vddk-libdir=/home/vddk8.0.1 -io  vddk-thumbprint=1B:83:D8:5A:33:31:62:DB:BA:9E:73:6D:A8:29:14:48:3F:82:F6:FD  -ip /home/passwd Auto-esx7.0-rhel9.1-only-root-luks-redhat123 --key "/dev/mapper/rhel_bootp--73--75--123-root":key:redhat123
[   0.2] Setting up the source: -i libvirt -ic vpx://administrator%40vsphere.local.213.93/data/10.73.212.38/?no_verify=1 -it vddk Auto-esx7.0-rhel9.1-only-root-luks-redhat123
[   2.0] Opening the source
[  10.1] Inspecting the source
[  23.0] Checking for sufficient free disk space in the guest
[  23.0] Converting Red Hat Enterprise Linux 9.1 Beta (Plow) to run on KVM
virt-v2v: This guest has virtio drivers installed.
[ 148.6] Mapping filesystem data to avoid copying unused and blank areas
virt-v2v: warning: fstrim on guest filesystem 
/dev/mapper/luks-cf19e4ac-3f54-450b-b801-72e539a8f870 failed.  Usually you 
can ignore this message.  To find out more read "Trimming" in virt-v2v(1).

Original message: fstrim: fstrim: /sysroot/: the discard operation is not 
supported
[ 150.1] Closing the overlay
[ 150.4] Assigning disks to buses
[ 150.4] Checking if the guest needs BIOS or UEFI to boot
[ 150.4] Setting up the destination: -o libvirt
[ 152.0] Copying disk 1/1
█ 100% [****************************************]
[ 517.9] Creating output metadata
[ 518.0] Finishing off

3.Check the guest after v2v conversion, checkpoints of guest are passed

Result:

  The bug has been fixed, move the bug from ON_QA to VERIFIED

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (virt-v2v bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2023:6376