[Bug 1092494] [NEW] Failing to mount rpc_pipefs

[gianluca]

Public bug reported:

I have a classroom of PCs with Ubuntu 12.04. Every PC mounts the
directory /home/users from an NFSv4 server using Kerberos
authentication. Most of the time everything works correctly, but it may
happen that the directory is not mounted. Upon inspection of the system,
I discovered that rpc.gssd and and rpc.statd are not running. They
cannot be started manually because the rpc_pipefs filesystem is not
mounted.

Doing a 'ps aux | grep pipefs' I have:

root       818  0.0  0.0   2720   256 ?        D    08:34   0:00 mount -t rpc_pipefs rpc_pipefs /run/rpc_pipefs
root       991  0.0  0.0   2720   256 ?        D    08:34   0:00 mount -t rpc_pipefs rpc_pipefs /run/rpc_pipefs

There are two instances of mount at the same time, and both failed.
Every attempt to manually mount rpc_pipefs fails in the same way. My
hypothesis is that there is a race condition between the upstart jobs
for gssd  and statd: both jobs try to mount rpc_pipefs, and if things go
badly, the two attempts happen concurrently, causing the failure.

This is actually pointed out in a comment to bug #811823:

https://bugs.launchpad.net/ubuntu/+source/nfs-
utils/+bug/811823/comments/6

** Affects: nfs-utils (Ubuntu)
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to nfs-utils in Ubuntu.
https://bugs.launchpad.net/bugs/1092494

Title:
  Failing to mount rpc_pipefs

Status in “nfs-utils” package in Ubuntu:
  New

Bug description:
  I have a classroom of PCs with Ubuntu 12.04. Every PC mounts the
  directory /home/users from an NFSv4 server using Kerberos
  authentication. Most of the time everything works correctly, but it
  may happen that the directory is not mounted. Upon inspection of the
  system, I discovered that rpc.gssd and and rpc.statd are not running.
  They cannot be started manually because the rpc_pipefs filesystem is
  not mounted.

  Doing a 'ps aux | grep pipefs' I have:

  root       818  0.0  0.0   2720   256 ?        D    08:34   0:00 mount -t rpc_pipefs rpc_pipefs /run/rpc_pipefs
  root       991  0.0  0.0   2720   256 ?        D    08:34   0:00 mount -t rpc_pipefs rpc_pipefs /run/rpc_pipefs

  There are two instances of mount at the same time, and both failed.
  Every attempt to manually mount rpc_pipefs fails in the same way. My
  hypothesis is that there is a race condition between the upstart jobs
  for gssd  and statd: both jobs try to mount rpc_pipefs, and if things
  go badly, the two attempts happen concurrently, causing the failure.

  This is actually pointed out in a comment to bug #811823:

  https://bugs.launchpad.net/ubuntu/+source/nfs-
  utils/+bug/811823/comments/6

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nfs-utils/+bug/1092494/+subscriptions