[Bug 2102236] Re: plucky/s390x (with root disk on dm?) does not come up after kernel update + reboot

[Frank Heimes]

I think this could be one way to solve this:

diff -U 3 ~/udev_orig /usr/share/initramfs-tools/hooks/udev
--- /home/ubuntu/udev_orig	2025-03-19 12:36:21.643873203 +0000
+++ /usr/share/initramfs-tools/hooks/udev	2025-03-19 12:59:12.031514341 +0000
@@ -73,7 +73,10 @@
   # Skip rules generated by chzdev as they are handled by its own hook.
   # See LP: #2044104
   if [ $ZDEV_FILTERING -eq 1 ] && /usr/sbin/chzdev --is-owner "$rules"; then
-    continue;
+    # only skip if rule is explicitly NOT configured as zdev_early
+    if [ $(grep "ENV{zdev_early}=\"0\"" "$rules") ]; then
+      continue;
+    fi
   fi
 
   if [ ! -e "/usr/lib/${rules#/etc/}" ]; then

In other words (and iianm):
- if rule is configured as zdev_early=1, it's explicitly asked to be added to the initrd
- if rule is configured as zdev_early=0, it's explicitly asked to NOT be added to the initrd
- if zdev_early is not (explicitly) set at all, it's safe(r) to be copied to initrd (this is the usual no autoconfig case on non-DPM systems before s390-tools v2.33.0 / before oracular)

but needs review

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to initramfs-tools in Ubuntu.
https://bugs.launchpad.net/bugs/2102236

Title:
  plucky/s390x (with root disk on dm?) does not come up after kernel
  update + reboot

Status in Ubuntu on IBM z Systems:
  New
Status in initramfs-tools package in Ubuntu:
  New
Status in linux package in Ubuntu:
  Invalid

Bug description:
  While using the plucky daily from March 12 (that still comes with kernel 6.12)
  (and working around LP#2101831, by forcing the installation to not apply any updates)
  I get a system installed, which is at kernel level 6.12.

  Since 6.14 is out (in plucky release) since yesterday (March 12th) I tried to upgrade from 6.12 to 6.14,
  and the update itself seemed to be smooth (I couldn't find any errors while doing a full-upgrade in the terminal - see attached logs).

  But after executing a reboot, the system (in 3 different
  configurations, that use dm) does not come up again, and ends up in
  busybox, complaining that the root device couldn't be found:

  # s1lp15 FCP/SCSI multipath with LVM
  ALERT!  /dev/disk/by-id/dm-uuid-LVM-SlleSC5YA825VOM3t0KHBVFrLJcNWsnwZsObNziIB9Bk2mSVphnuTEOQ2eFiBbE1 does not exist.  Dropping to a shell!

  # s1lp15 2DASDs with LVM:
  ALERT!  /dev/disk/by-id/dm-uuid-LVM-ePTbsojYPfgMacKXwpIMNMvxk80qGzlPhRYw7DJlovmqHyla9TK6NGc70p1JN29b does not exist.  Dropping to a shell!

  # s1lp15 FCP/SCSI Multipath no LVM
  ALERT!  /dev/disk/by-id/dm-uuid-part1-mpath-36005076306ffd6b60000000000002603 does not exist.  Dropping to a shell!

  However, using a single disk without dm (so: no multipath, no lvm) the
  system is able to come up again after the reboot (after a kernel
  upgrade).

  # s1lp15 single DASD no LVM
  here the root device is:
  root=/dev/disk/by-path/ccw-0.0.260b-part1
  and it exists.

  In the 3 different cases that (that repeatedly fail) "/dev/disk/by-id"
  is missing.

  I am not sure yet what's causing this,
  it can be an issue with the device-mapper/lvm2
  but also udev rules or kernel.
  So maybe I need to add the kernel as affected component too (for now).

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu-z-systems/+bug/2102236/+subscriptions