On Fri, 2008-04-18 at 09:39 -0700, Rajouri Jammu wrote:
> Hi,
> 
> 
> I'm trying to bring up a stacked file system on top of an NFS mount
> point both via automount.
> 
> 
> For example:
> On my client system I have two directories /data1 and /data2
> /data1/nfs_01
> /data2/nfs_01
> 
> 
> nfs_server:/data/nfs_01 is mounted on /data1/nfs_01 (type nfs)
> 
> 
> /data1/nfs_01 is mounted on /data2/nfs_01
> 
> 
> The automount config files look like:
> /etc/auto.data1:
> nfs_01   nfs_server:/data/nfs_01
> 
> 
> 
> /etc/auto.data2:
> nfs_01   -fstype=ecryptfs :/data1/nfs_01
> 
> 
> 
> 
> The problem:
> 
> 
> When I try to access /data2/nfs_01 it kicks in automount to
> mount /data1/nfs_01, which in turn kicks in automount again to mount
> nfs_server:/data/nfs_01 on /data1/nfs_01.
> 
> 
> The second mount waits for 10secs for a lock file held by the first
> mount in mount_nfs.c after mkdir_path().

So you're running version 4?

> 
> 
> Is there a way I can configure things such that the second mount
> doesn't wait?

There is a patch around to add a configure option
"--disable-mount-locking" but you need to be sure you're mount(8) is at
least reasonably up to date.

You may want to look in
http://www.kernel.org/pub/linux/daemons/autofs/v4/

> 
> 
> 
> The lock file is a single global.
> 
> Should the file be specific to the mount point?

No, it is meant help prevent corruption of /etc/mtab as the locking in
mount(8) was broken for a long time.

Ian



_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs

Reply via email to