Hi,

Those affected by this issue might want to check if they happen to have 
a NFS loopback mount:

% mount | egrep "^`hostname`.*type nfs"

should not report a single line.

I've been struck by this error 4-5 times a day on a single server until 
I found out that automount resources my host offered within our network 
did not get mounted locally using the "--bind" mount option, but using 
mount type NFS. As posted several times on the kernel mailing list a 
'mount -t nfs localhost:' is in general a bad idea (see e.g. 
http://kerneltrap.org/mailarchive/linux-kernel/2007/9/21/271649/thread)

Why did my Ubuntu 10.04 installation use a NFS loopback mount? Well, 
the first few lines of /etc/hosts on a freshly installed Ubuntu 10.04 read

   127.0.0.1       localhost
   127.0.1.1       NAME_YOU_USE_FOR_YOUR_MACHINE
   
   # The following lines are desirable for IPv6 capable hosts
   ::1     localhost ip6-localhost ip6-loopback

Your automount configuration is probably set up such that /etc/auto.home 
(or the equivalent LDAP resource) provides a setting similar to the 
following:

  NAME_YOU_USE_FOR_YOUR_MACHINE  NAME_YOU_USE_FOR_YOUR_MACHINE:/path/to
/some-directory

When trying to access /NAME_YOU_USE_FOR_YOUR_MACHINE/some-directory 
while logged in to NAME_YOU_USE_FOR_YOUR_MACHINE the ip address of 
NAME_YOU_USE_FOR_YOUR_MACHINE is resolved to 127.0.1.1 which is neither 
the ip address of your loopback device (127.0.0.1) nor of any of 
your network cards. That seems to be why automount decides to do a NFS 
loopback mount, not a "--bind" mount.

According to Section 10.4 of the Debian Reference Manual the address 
127.0.1.1 is only listed in /etc/hosts because "[s]ome software (e.g., 
GNOME) expects the system hostname to be resolvable to an IP address 
with a canonical fully qualified domain name. This is really improper 
because system hostnames and domain names are two very different things; 
but there you have it. In order to support that software, it is 
necessary to ensure that the system hostname can be resolved. Most often 
this is done by putting a line in /etc/hosts containing some IP address 
and the system hostname. If your system has a permanent IP address then 
use that; otherwise use the address 127.0.1.1". (citated from 
http://qref.sourceforge.net/quick/ch-gateway.en.html) 

If your host is affected by this issue and is offering automount 
resources, it can be safely assumed to have (at least one) permanent IP 
address. Which implicates that the line with the ip address 127.0.1.1 in 
/etc/hosts can be removed/commented out/replaced with the proper ip 
address (one of) your network card(s) uses. No NFS loopback mounts will 
happen any more then, but decent "--bind" mounts such that the problem 
described in the postings of this thread will not be triggered any more. 
At least that's what solved it for 50+ hosts running Ubuntu 10.04 24/7 
in our network, not a single one has crashed within weeks since having 
applied this modification while all of them provide automount resources 
used both remotely and locally.

Hope this helps.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/658131

Title:
  Overload the server ... __mutex_lock_slowpath

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to