U1604: Installing root (/) into an Ext4 formatted LUKS partition creates /dev/dm-0, rather than /dev/mapper/sdaX_crypt

Bug #1531387 reported by GizmoChicken
14
This bug affects 3 people
Affects Status Importance Assigned to Milestone
ubiquity (Ubuntu)
Fix Released
Low
Unassigned

Bug Description

Under Ubuntu 16.04, installing root (/) into an Ext4 formatted LUKS partition creates /dev/dm-0, rather than /dev/mapper/sdaX_crypt.

On the other hand, installing root (/) into a Btrfs formatted LUKS partition creates /dev/mapper/sdaX_crypt, rather than /dev/dm-0.

I'm under the impression that it is more appropriate to create /dev/mapper/sdaX_crypt, rather than /dev/dm-0, when installing root (/) into a LUKS partition. See https://www.redhat.com/archives/linux-lvm/2009-August/msg00010.html

Whether this raises to the level of a bug, I'm not sure. But it certainly is peculiar that ubiquity performs correctly when installing root (/) into a Btrfs formatted LUKS partition, but performs incorrectly when installing root (/) into an Ext4 formatted LUKS partition. If truly a bug, I hope that it will be fixed before the release of U1604.

Also, this bug may be related to Bug #1523194.

Tags: wily xenial
Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in ubiquity (Ubuntu):
status: New → Confirmed
Revision history for this message
Phillip Susi (psusi) wrote :

Both files should be created with one being a symbolic link to the other. What is the actual problem that you experience? Does the install not boot?

Changed in ubiquity (Ubuntu):
status: Confirmed → Incomplete
Revision history for this message
GizmoChicken (gizmochicken) wrote :

It boots. And like I said, I'm not sure if this constitutes a full bug. But it certainly is an inconsistency.

That is:

1) Installing root (/) into an Ext4 formatted LUKS partition creates /dev/dm-0, rather than /dev/mapper/sdaX_crypt.

2) Installing root (/) into a Btrfs formatted LUKS partition creates /dev/mapper/sdaX_crypt, rather than /dev/dm-0.

The above doesn't seem odd?

And as I wrote before, I'm under the impression that it is more appropriate to create /dev/mapper/sdaX_crypt, rather than /dev/dm-0, when installing root (/) into a LUKS partition. See https://www.redhat.com/archives/linux-lvm/2009-August/msg00010.html

The information at the link is inaccurate?

Changed in ubiquity (Ubuntu):
status: Incomplete → Confirmed
tags: added: wily xenial
Changed in ubiquity (Ubuntu):
importance: Undecided → High
Revision history for this message
Phillip Susi (psusi) wrote :

As I said before, both should be created. The /dev/mapper one is the one you should normally use, but /dev/dm-x also should exist.

Changed in ubiquity (Ubuntu):
importance: High → Low
Revision history for this message
Phillip Susi (psusi) wrote :

I'm not able to reproduce this. I've tried installing with both ext4 and btrfs and in both cases I get /dev/mapper/xvda_crypt as a symlink to /dev/dm-0, which is the actual block device. Can you provide more details?

Changed in ubiquity (Ubuntu):
status: Confirmed → Incomplete
Revision history for this message
GizmoChicken (gizmochicken) wrote :

Phillip,

Thanks much for the comment. Also, thanks for you involvement in Bug #1523194. I originally reported the current bug/observation because I thought maybe it was related to Bug #1523194, which at least for me, is a fairly important bug. Now, I doubt that there is a relationship, and so I'm less interested in the bug/observation reported here. And so, at least for now, I agree with your re-characterization of importance: High → Low.

Nonetheless, I continue to believe that something odd is going on here. But until I spend a little more time with this, I'm not quite sure how to explain the situation better.

How about we leave this bug marked "incomplete" for now, and I'll get back to you in a few days with more information and a better explanation of what I suspect is going wrong?

Best regards,
GizmoChicken

Revision history for this message
Phillip Susi (psusi) wrote :

Sure -- that's what incomplete is for: waiting for more information.

Revision history for this message
GizmoChicken (gizmochicken) wrote :

Looks like this bug/inconsistency has been fixed, at least for the time being.

Just to clarify my previous observation…

Prior to the fix, when installing root (/) into an Ext4 formatted LUKS partition, the system recognized the LUKS device as /dev/dm-0, rather than as /dev/mapper/sdaX_crypt. The are various ways that this could be observed. However, the easiest way to see this is by looking at sysmon. See U1604-2016-02-03.png, which is a screenshot showing sysmon running on a system installed using the U1604 daily build from 2016-02-03.

And here’s what can be observed after installing with a more recent daily build…

After the fix, when installing root (/) into an Ext4 formatted LUKS partition, the system recognizes the LUKS device as /dev/mapper/sdaX_crypt. See U1604-2016-02-26.png, which is a screenshot showing sysmon running on a system installed using the U1604 daily build from 2016-02-26.

I’ll just add…

This bug/inconsistency seems to have been unique to installing root (/) into an Ext4 formatted LUKS partitions. When installing root (/) into a BTRFS formatted LUKS partition, the system recognized the LUKS device as /dev/mapper/sdaX_crypt, both before and after the fix. Similarly, when installing root (/) into an XFS formatted LUKS partition, the system recognized the LUKS device as /dev/mapper/sdaX_crypt, both before and after the fix.

Don’t know who submitted the fix, or if this report had anything to do with it, but thanks!

Revision history for this message
GizmoChicken (gizmochicken) wrote :

U1604-2016-02-26.png is a screenshot showing sysmon running on a system installed using the U1604 daily build from 2016-02-26.

Revision history for this message
GizmoChicken (gizmochicken) wrote :

U1604-2016-02-03.png is a screenshot showing sysmon running on a system installed using the U1604 daily build from 2016-02-03.

Changed in ubiquity (Ubuntu):
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.