lubuntu hirsute install fail mkfs.xfs

Bug #1917724 reported by Chris Guiver
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
calamares (Ubuntu)
New
Undecided
Unassigned

Bug Description

Lubuntu hirsute QA-test install (daily) on
-dell [optiplex] 755 (c2d-e6850, 5gb, amd/ati radeon rv516/x1300/x1550)

Testcase: re-use /home (sda8) & format of sda9 for / (XFS this time, it was reiserfs; but I've been alternating between XFS & BTRFS usually)

The installer failed to create partition on disk 'ST3160815AS'.
==========================================================================================
Create a new partition (22.00 GiB, xfs) on ‘/dev/sda’
==========================================================================================
==========================================================================================
Job: Create new partition on device ‘/dev/sda’
==========================================================================================
Command: mkfs.xfs -f /dev/sda9
==========================================================================================

Note: this was a repeat install of lp 1864787 / 1914521
so the problem could be related to those.. error listed differs though

In session.log is "Command: sfdisk --force --append /dev/sda" so I'm pretty sure this is a duplicate of https://bugs.launchpad.net/ubuntu/+source/calamares/+bug/1864787

ProblemType: Bug
DistroRelease: Ubuntu 21.04
Package: calamares 3.2.37-0ubuntu1
ProcVersionSignature: Ubuntu 5.10.0-14.15-generic 5.10.11
Uname: Linux 5.10.0-14-generic x86_64
.etc.calamares.modules.finished.conf:
 ---
 restartNowMode: user-checked
 restartNowCommand: "systemctl -i reboot"
.etc.calamares.modules.shellprocess_logs.conf:
 ---
 dontChroot: true
 timeout: 30
 script:
     - calamares-logs-helper @@ROOT@@
.etc.calamares.modules.unpackfs.conf:
 ---
 unpack:
     - source: "/cdrom/casper/filesystem.squashfs"
         sourcefs: "squashfs"
         destination: ""
ApportVersion: 2.20.11-0ubuntu59
Architecture: amd64
CasperMD5CheckResult: unknown
CasperVersion: 1.458
CurrentDesktop: LXQt
Date: Thu Mar 4 18:34:54 2021
LiveMediaBuild: Lubuntu 21.04 "Hirsute Hippo" - Alpha amd64 (20210303)
RelatedPackageVersions:
 calamares-settings-ubuntu-common 1:21.04.2
 calamares-settings-lubuntu 1:21.04.2
 xfsprogs 5.10.0-2ubuntu3
 btrfs-progs 5.10.1-1build1
SourcePackage: calamares
UpgradeStatus: No upgrade log present (probably fresh install)

Revision history for this message
Chris Guiver (guiverc) wrote :
Revision history for this message
Ubuntu QA Website (ubuntuqa) wrote :

This bug has been reported on the Ubuntu ISO testing tracker.

A list of all reports related to this bug can be found here:
http://iso.qa.ubuntu.com/qatracker/reports/bugs/1917724

tags: added: iso-testing
Chris Guiver (guiverc)
description: updated
Revision history for this message
Chris Guiver (guiverc) wrote :

http://iso.qa.ubuntu.com/qatracker/milestones/419/builds/227564/testcases/1701/results/ is link to is today's notes

After the install, I had issues. On reboot of daily ISO, the / partition was UNKNOWN which was expected.

I loaded up KDE Partition Manager, deleted the UNKNOWN partition (good) then formatted it REISERfs (habit; it's easy to spot), but that failed in what looked like the same error as this install? (mkfs.reiserfs)...

then KDE Partition Manager crashed????

Usually those occur without issue.. I exit KDE Partition Manager & attempt QA-test install again... not today.

Crash of KDE Partition Manager is 1917753 (currently marked private? it's a live session so can't imagine there is anything private there)

I still believe this is a repeat of 1864787, except
- on screen error message looked different (session.log was closer)
- reboot into live & KDE partition manager acted differently giving a like error?? then it crashed??

it feels different, I'm too tired now to repeat the testing.

Revision history for this message
ԜаӀtеr Ⅼарсһуnѕkі (wxl) wrote :

I have a wee bit of a concern that the kpmcore patch we added to get EFI installs working in Calamares may have had an impact here. I hope not, but I'm concerned. If that's the case, we may have to throw the latest Git master of kpmcore in a PPA and rebuild Calamares against it and try with that. In which case we may have to either drop that patch and issue release notes with a workaround for the label issue or ship Git master (probably not).

To reduce noise, can we just forget about reusing and extra partitions and stuff and simply do a fresh partition table with either xfs, reiserfs, or btrfs (i.e. 3 separate tests) and see if it fails similarly? In other words, I want to test just whether or not creating the file system on a fresh table works. This should just work. It certainly should just work in kpm, as it has in the past. If those fail and especially if they all fail, that's trouble.

Changed in calamares (Ubuntu):
status: New → Incomplete
importance: Undecided → Critical
Revision history for this message
Leó Kolbeinsson (leok) wrote :

Tested installing on clean partitions 3 seperate tests- xfs,btrfs and reiserfs - all installs without partitioning problems - see results on the testing tracker here:

http://iso.qa.ubuntu.com/qatracker/milestones/419/builds/227592/testcases/1701/results

Revision history for this message
ԜаӀtеr Ⅼарсһуnѕkі (wxl) wrote :

Ok so now I'm less concerned, let's try to figure out how to actually reproduce this. Clearly an existing partition table is necessary.

Changed in calamares (Ubuntu):
importance: Critical → Undecided
Revision history for this message
Chris Guiver (guiverc) wrote :

Yesterday may have just been unlucky, after all the issues with a cleanly booted daily and KDE Partition Manager crashing (https://bugs.launchpad.net/ubuntu/+source/partitionmanager/+bug/1917753) have NOT occurred today.

I was able to use KDE Partition Manager today as I normally do, to setup the box and do an install.

Yes if I attempted to 'format' the partition with calamares it crashed with this error (tried once today; failed), but if calamares does not format the partition (tried once today) it worked, ie. workaround today worked (where yesterday it failed).

I'll continue some installs following Walter's #4 comment (install occurring currently on dc7700; and I'll move to d755-8 next; as I've only had this (or related 1864787) issues before on d755-5 or d755-8; ie. a dell optiplex model 755 though differing vintages) but I won't report back unless issues occur.

Chris Guiver (guiverc)
Changed in calamares (Ubuntu):
status: Incomplete → New
Revision history for this message
Chris Guiver (guiverc) wrote :

Repeat QA-test install on d755-5 (box mentioned in this report)
re-use sda8 as /home, format sda9 as XFS

Installation Failed

command mkfs.xfs -f /dev/sda9

I've changed status back to NEW, but

I am CONVINCED this issue relates to 1864787 & 1914521 and is likely(99.9%) a DUPLICATE (just a different message)

Could the message have just changed? maybe I get this message when I use XFS, but other messages for other file-systems?, ie. identical cause/issue/bug just message varies on what parameters I do that triggers the cause (I'm tagging bug by message on screen!)

To post a comment you must log in.