[Bug 1653936] [NEW] bogus nilfs2 partition detection breaks /dev/disk/by-partuuid creation

[Matthew Vernon]

Public bug reported:

When my system booted, it failed to create a /dev/disk/by-partuuid link
for one partition (/dev/nvme0n1p11) that means that ceph is unable to
start up (the relevant osd looks for /var/lib/ceph/osd/ceph-388/journal
which is a symlink to /dev/disk/by-
partuuid/d2ace848-7e2d-4395-a195-a4428631b333 which doesn't exist). See
the shell outputs below:

root at sto-3-3:~# file /var/lib/ceph/osd/ceph-388/journal
/var/lib/ceph/osd/ceph-388/journal: broken symbolic link to /dev/disk/by-partuuid/d2ace848-7e2d-4395-a195-a4428631b333
root at sto-3-3:~# blkid /dev/nvme0n1p11 
/dev/nvme0n1p11: PARTLABEL="ceph journal" PARTUUID="d2ace848-7e2d-4395-a195-a4428631b333"
root at sto-3-3:~# udevadm info /dev/nvme0n1p11
P: /devices/pci0000:00/0000:00:02.0/0000:02:00.0/nvme/nvme0/nvme0n1/nvme0n1p11
N: nvme0n1p11
E: DEVNAME=/dev/nvme0n1p11
E: DEVPATH=/devices/pci0000:00/0000:00:02.0/0000:02:00.0/nvme/nvme0/nvme0n1/nvme0n1p11
E: DEVTYPE=partition
E: ID_PART_TABLE_TYPE=gpt
E: ID_PART_TABLE_UUID=8c11f541-fdc1-4299-923b-dc57d8a38619
E: MAJOR=259
E: MINOR=11
E: SUBSYSTEM=block
E: TAGS=:systemd:
E: USEC_INITIALIZED=7070417

root at sto-3-3:~# grep nvme0n1p11 /var/log/syslog
Jan  4 09:17:20 sto-3-3 systemd-udevd[106928]: incorrect nilfs2 checksum on /dev/nvme0n1p11

...so the bogus nilfs2 checksum failure is causing udev to not create
the ID_PART_ENTRY_UUID and similar entries (and hence the /dev/disk/by-
partuuid link) and is breaking my ceph setup.

I think this is related to this upstream issue:
https://github.com/karelzak/util-linux/issues/361

...this is quite a serious bug, as it's causing unrelated software
(ceph) to break.

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: util-linux 2.27.1-6ubuntu3.1
ProcVersionSignature: Ubuntu 4.4.0-47.68-generic 4.4.24
Uname: Linux 4.4.0-47-generic x86_64
ApportVersion: 2.20.1-0ubuntu2.4
Architecture: amd64
Date: Wed Jan  4 10:19:19 2017
InstallationDate: Installed on 2016-11-01 (63 days ago)
InstallationMedia: Ubuntu-Server 16.04.1 LTS "Xenial Xerus" - Release amd64 (20160719)
SourcePackage: util-linux
UpgradeStatus: No upgrade log present (probably fresh install)

** Affects: util-linux (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: amd64 apport-bug xenial

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to util-linux in Ubuntu.
https://bugs.launchpad.net/bugs/1653936

Title:
  bogus nilfs2 partition detection breaks /dev/disk/by-partuuid creation

Status in util-linux package in Ubuntu:
  New

Bug description:
  When my system booted, it failed to create a /dev/disk/by-partuuid
  link for one partition (/dev/nvme0n1p11) that means that ceph is
  unable to start up (the relevant osd looks for
  /var/lib/ceph/osd/ceph-388/journal which is a symlink to /dev/disk/by-
  partuuid/d2ace848-7e2d-4395-a195-a4428631b333 which doesn't exist).
  See the shell outputs below:

  root at sto-3-3:~# file /var/lib/ceph/osd/ceph-388/journal
  /var/lib/ceph/osd/ceph-388/journal: broken symbolic link to /dev/disk/by-partuuid/d2ace848-7e2d-4395-a195-a4428631b333
  root at sto-3-3:~# blkid /dev/nvme0n1p11 
  /dev/nvme0n1p11: PARTLABEL="ceph journal" PARTUUID="d2ace848-7e2d-4395-a195-a4428631b333"
  root at sto-3-3:~# udevadm info /dev/nvme0n1p11
  P: /devices/pci0000:00/0000:00:02.0/0000:02:00.0/nvme/nvme0/nvme0n1/nvme0n1p11
  N: nvme0n1p11
  E: DEVNAME=/dev/nvme0n1p11
  E: DEVPATH=/devices/pci0000:00/0000:00:02.0/0000:02:00.0/nvme/nvme0/nvme0n1/nvme0n1p11
  E: DEVTYPE=partition
  E: ID_PART_TABLE_TYPE=gpt
  E: ID_PART_TABLE_UUID=8c11f541-fdc1-4299-923b-dc57d8a38619
  E: MAJOR=259
  E: MINOR=11
  E: SUBSYSTEM=block
  E: TAGS=:systemd:
  E: USEC_INITIALIZED=7070417

  root at sto-3-3:~# grep nvme0n1p11 /var/log/syslog
  Jan  4 09:17:20 sto-3-3 systemd-udevd[106928]: incorrect nilfs2 checksum on /dev/nvme0n1p11

  ...so the bogus nilfs2 checksum failure is causing udev to not create
  the ID_PART_ENTRY_UUID and similar entries (and hence the /dev/disk
  /by-partuuid link) and is breaking my ceph setup.

  I think this is related to this upstream issue:
  https://github.com/karelzak/util-linux/issues/361

  ...this is quite a serious bug, as it's causing unrelated software
  (ceph) to break.

  ProblemType: Bug
  DistroRelease: Ubuntu 16.04
  Package: util-linux 2.27.1-6ubuntu3.1
  ProcVersionSignature: Ubuntu 4.4.0-47.68-generic 4.4.24
  Uname: Linux 4.4.0-47-generic x86_64
  ApportVersion: 2.20.1-0ubuntu2.4
  Architecture: amd64
  Date: Wed Jan  4 10:19:19 2017
  InstallationDate: Installed on 2016-11-01 (63 days ago)
  InstallationMedia: Ubuntu-Server 16.04.1 LTS "Xenial Xerus" - Release amd64 (20160719)
  SourcePackage: util-linux
  UpgradeStatus: No upgrade log present (probably fresh install)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/1653936/+subscriptions