[Bug 696435] Re: wait-for-root fails to detect nbd root

[Alkis Georgopoulos]

I'm the bug reporter, and I have a problem in doing the verification.

a) The initial testcase that I reported happens in both 4.4 unpatched
and 4.10. I.e. the bug that I reported is not yet fixed.

b) Some Ubuntu developer wrote a new testcase as part of doing the SRU.
I cannot reproduce that testcase neither in 4.4 unpatched nor in 4.10,
i.e. `udevadm monitor` does show add/change events for nbd devices for
me in both kernels. I.e. that sounds like a different bug that I never
saw.

So I'm not sure how I can help here...

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/696435

Title:
  wait-for-root fails to detect nbd root

Status in linux package in Ubuntu:
  Fix Released
Status in nbd package in Ubuntu:
  Invalid
Status in systemd package in Ubuntu:
  Fix Released
Status in linux source package in Xenial:
  Fix Committed
Status in systemd source package in Xenial:
  Confirmed

Bug description:
  [Impact]
  Kernel does not generate any events when ndb-client connects /dev/nbd0 devices, therefore it is impossible to monitor/react to the state of /dev/nbd0.

  [Fix]
  Generate change uevent when size of /dev/nbd0 changes

  [Testcase]
  * Start udevadm monitor
  * modprobe nbd
  * use ndb-client to connect something to /dev/nbd0
  * observe that there are change udev events generated on /dev/nbd0 itself

  [Regression Potential]
  There is no change to existing uevents, or their ordering.
  There is now an addition change event which will cause systemd to mark ndb devices as ready and trigger appropriate actions

  [Original Bug Report]

  When using an nbd root, wait-for-root blocks for 30 seconds before
  booting continues successfully.

  Using Ubuntu Natty, related packages versions:
      nbd-client 1:2.9.16-6ubuntu1
      initramfs-tools 0.98.1ubuntu9

  The wait-for-root call from /usr/share/initramfs-tools/scripts/local:
   while [ -z "${FSTYPE}" ]; do
    FSTYPE=$(wait-for-root "${ROOT}" ${ROOTDELAY:-30})

    # Run failure hooks, hoping one of them can fix up the system
    # and we can restart the wait loop.  If they all fail, abort
    # and move on to the panic handler and shell.
    if [ -z "${FSTYPE}" ] && ! try_failure_hooks; then
     break
    fi
   done

  I replaced wait-for-root with a sh script that did `set >&2`, here are the relevant environment variables at the time wait-for-root was called:
  ROOT='/dev/nbd0'
  ROOTDELAY=''
  ROOTFLAGS=''
  ROOTFSTYPE=''
  nbdroot='192.168.0.1,2011'

  It's probably worth noting that "nbd0: unknown partition table" was
  displayed asynchronously 1-2 seconds after wait-for-root was invoked
  and while it was still waiting. But I tried adding a "sleep 5" as the
  last line of local-top/nbd, so that the nbd message was displayed a
  lot before wait-for-root was called, and it didn't make a difference.
  So I don't think a race condition is involved in this problem.

  Temporarily I'm passing rootdelay=1 in the kernel command line to work
  around the problem.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/696435/+subscriptions