Bug 751731

I worked out this was one of a couple of issues I was seeing with F16 as well. The patch from bug 747020 works for me if I work around the other issue.

+++ This bug was initially created as a clone of Bug #747020 +++

When we updated from autofs-5.0.1-0.rc2.156.el5_7.1 to autofs-5.0.1-0.rc2.156.el5_7.3 some of our mounts stopped working. Our relevant set up is
that the file /etc/auto.master contains the line
+auto.master
and in the auto.master map, one of the mounts that breaks has key /home and value
/etc/auto.home            -rw,intr,noquota,noac,actimeo=0
The file /etc/auto.home contains the line
+auto.home
and a failing mount in this has a key like servername and value like
-rw,intr,quota,noac,actimeo=0 /mount1 &:/path/mount1 /mount2 &:/path/mount2 /mount3 &:/path/mount3 /mount4 &:/path/mount4 /mount5 &/path/mount5
If I downgrade to autofs-5.0.1-0.rc2.156.el5_7.1 they start working again.

--- Additional comment from ikent on 2011-10-19 00:19:31 EDT ---

(In reply to comment #0)
> When we updated from autofs-5.0.1-0.rc2.156.el5_7.1 to
> autofs-5.0.1-0.rc2.156.el5_7.3 some of our mounts stopped working. Our relevant
> set up is
> that the file /etc/auto.master contains the line
> +auto.master
> and in the auto.master map, one of the mounts that breaks has key /home and
> value
> /etc/auto.home            -rw,intr,noquota,noac,actimeo=0
> The file /etc/auto.home contains the line
> +auto.home
> and a failing mount in this has a key like servername and value like
> -rw,intr,quota,noac,actimeo=0 /mount1 &:/path/mount1 /mount2 &:/path/mount2
> /mount3 &:/path/mount3 /mount4 &:/path/mount4 /mount5 &/path/mount5
> If I downgrade to autofs-5.0.1-0.rc2.156.el5_7.1 they start working again.

I'm having difficulty understanding what these map entries
look like, can you post the maps please.

--- Additional comment from ikent on 2011-10-19 04:13:42 EDT ---

I've been able to reproduce this.
Will continue to work on it.

--- Additional comment from ikent on 2011-10-20 00:08:30 EDT ---

I have to ask, why has this been logged directly in Bugzilla?
To make the correct people aware of this regression it needs to
be logged via the normal support channel.

--- Additional comment from m.a.young.uk on 2011-10-20 05:10:32 EDT ---

(In reply to comment #3)
> I have to ask, why has this been logged directly in Bugzilla?
> To make the correct people aware of this regression it needs to
> be logged via the normal support channel.

Our RHEL boxes are currently on academic licenses, which gives us package updates but is "self-supporting", so I don't think we have any access to the support channels.

--- Additional comment from ikent on 2011-10-20 11:00:45 EDT ---

(In reply to comment #4)
> (In reply to comment #3)
> > I have to ask, why has this been logged directly in Bugzilla?
> > To make the correct people aware of this regression it needs to
> > be logged via the normal support channel.
> 
> Our RHEL boxes are currently on academic licenses, which gives us package
> updates but is "self-supporting", so I don't think we have any access to the
> support channels.

OK, I'll see if I can work out who to ping to get this the
attention it deserves.

--- Additional comment from ikent on 2011-10-24 22:02:22 EDT ---

A package with the above patch has been built and is
available at:
http://people.redhat.com/~ikent/autofs-5.0.1-0.rc2.160.bz747020.1

This may resolve the problem seen here, please test and report
your result.

--- Additional comment from ikent on 2011-10-24 22:03:51 EDT ---

Created attachment 529986 [details]
Patch - fix fix map source check in file lookup

--- Additional comment from m.a.young.uk on 2011-10-25 04:59:53 EDT ---

(In reply to comment #7)
> A package with the above patch has been built and is
> available at:
> http://people.redhat.com/~ikent/autofs-5.0.1-0.rc2.160.bz747020.1
> 
> This may resolve the problem seen here, please test and report
> your result.

I am testing that package and it works as expected; the problem mounts work and I haven't spotted any other problems.