This Debian bug report includes an exportfs problem which I suspect is
due to one export tree under the other, and an error message complaint
which I think is justified....
----- Forwarded message from Phil Edwards <[email protected]> -----
Subject: Bug#239158: Strange exportfs problems, fantastically unhelpful error message
From: Phil Edwards <[email protected]>
To: Debian Bug Tracking System <[email protected]>
Date: Sun, 21 Mar 2004 00:37:43 -0500
Package: nfs-kernel-server
Version: 1:1.0.6-1
Severity: normal
[~]# cat /etc/exports
/usr/local fenric(ro,async) hreidmar(ro,async)
/home fenric(rw,sync) hreidmar(rw,sync)
[~]# exportfs -v -r
exporting hreidmar:/usr/local
exporting fenric:/usr/local
exporting hreidmar:/home
exporting fenric:/home
exporting hreidmar:/usr/local to kernel
hreidmar:/usr/local: Invalid argument
exporting fenric:/usr/local to kernel
fenric:/usr/local: Invalid argument
reexporting hreidmar:/home to kernel
reexporting fenric:/home to kernel
[~]#
I have no idea why the export fails. If either line is commented, then
the export succeeds; presumably that's strong evidence of /something/,
but I don't know the nfs-kernel-server code at all and so can't say what.
My real complaint is that "Invalid argument" doesn't tell me a damn thing,
so I have no idea what to try next.
-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux fenric 2.4.22-fenric3 #1 SMP Sun Oct 12 00:45:53 EDT 2003 i686
Locale: LANG=C, LC_CTYPE=C
Versions of packages nfs-kernel-server depends on:
ii debconf 1.4.7 Debian configuration management sy
ii libc6 2.3.2.ds1-11 GNU C Library: Shared libraries an
ii libwrap0 7.6-ipv6.1-3 Wietse Venema's TCP wrappers libra
ii nfs-common 1:1.0.6-1 NFS support files common to client
ii sysvinit 2.85-9 System-V like init.
-- debconf information:
nfs-kernel-server/sync-default:
nfs-kernel-server/tcpwrappers-mountd:
----- End forwarded message -----
--
Chip Salzenberg - a.k.a. - <[email protected]>
"I wanted to play hopscotch with the impenetrable mystery of existence,
but he stepped in a wormhole and had to go in early." // MST3K
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
On Sun, Mar 21, 2004 at 10:35:15AM -0500, Chip Salzenberg wrote:
> This Debian bug report includes an exportfs problem which I suspect is
> due to one export tree under the other, and an error message complaint
> which I think is justified....
Well, what he tries to do is invalid from the kernel point of view.
And unfortunately there's no "EBADNFSEXPORTPLEASEREADTHEMANPAGE" :-)
I agree though this issue is confusing to many people. Wasn't there
some work going on to allow an arbitrary configuration of exports?
For each file handle, we need to establish the dentry chain anyway,
so it should not be too hard to find the proper export entry.
Olaf
--
Olaf Kirch | Stop wasting entropy - start using predictable
[email protected] | tempfile names today!
---------------+
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Chip Salzenberg wrote:
>
> This Debian bug report includes an exportfs problem which I suspect is
> due to one export tree under the other,
Possibly, if one of /usr/local or /home is a symlink to somewhere
underneath the other.
> and an error message complaint
> which I think is justified....
I second this complaint. On other OSes that case is checked in
userspace early and a sensible message printed, e.g.
irix# exportfs -av
re-exported /
exportfs: /opt: parent-directory (/) already exported
re-exported /mnt/scratch
> I have no idea why the export fails. If either line is commented, then
> the export succeeds; presumably that's strong evidence of /something/,
> but I don't know the nfs-kernel-server code at all and so can't say what.
You can get a better idea of what is failing by:
linux# echo 4 > /proc/sys/sunrpc/nfsd_debug
linux# exportfs -av
exporting foo.melbourne.sgi.com:/home/fmeh
exporting bar.melbourne.sgi.com:/home/fmeh
exporting foo.melbourne.sgi.com:/home
exporting bar.melbourne.sgi.com:/home
exporting foo.melbourne.sgi.com:/home/fmeh to kernel
exporting bar.melbourne.sgi.com:/home/fmeh to kernel
exporting foo.melbourne.sgi.com:/home to kernel
foo.melbourne.sgi.com:/home: Invalid argument
exporting bar.melbourne.sgi.com:/home to kernel
bar.melbourne.sgi.com:/home: Invalid argument
linux# tail /var/log/messages
...
Mar 22 09:49:53 4A:quux kernel: exp_export called for foo.melbourne.sgi.com:/home
(403/54526548 fl 24).
Mar 22 09:49:53 4A:quux kernel: exp_export: export not valid (Rule 3). <------
Mar 22 09:49:53 4A:quux kernel: exp_export called for bar.melbourne.sgi.com:/home
(403/54526548 fl 24).
Mar 22 09:49:53 4A:quux kernel: exp_export: export not valid (Rule 3). <------
The error message tells you...indirectly...what is going on. You need
to match it to the source of fs/nfsd/export.c:exp_export() to make any
sense of it. The "Rule 3" message happens when a child of the given
directory is already exported.
Greg.
--
Greg Banks, R&D Software Engineer, SGI Australian Software Group.
I don't speak for SGI.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs