[gentoo-user] mount.nfs stale nfs handle

[gentoo-user] mount.nfs stale nfs handle

[Alexander Puchmayr]

Hi there,

After upgrading my server to latest stable release of gentoo, none of my 
clients is able to mount any nfs share from the server anymore.

Symptoms:
$ mount -v -t nfs poseidon:/datadisk/ /mnt/gentoo/
mount.nfs: timeout set for Sun Jun 29 19:33:40 2014
mount.nfs: trying text-based options 
'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
mount.nfs: mount(2): Protocol not supported
mount.nfs: trying text-based options 'addr=192.168.1.6'
mount.nfs: prog 100003, trying vers=3, prot=6
mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
mount.nfs: prog 100005, trying vers=3, prot=17
mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
mount.nfs: mount(2): Stale NFS file handle
mount.nfs: trying text-based options 
'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
mount.nfs: mount(2): Protocol not supported
mount.nfs: trying text-based options 'addr=192.168.1.6'
mount.nfs: prog 100003, trying vers=3, prot=6
mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
mount.nfs: prog 100005, trying vers=3, prot=17
mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
mount.nfs: mount(2): Stale NFS file handle
[...]
mount.nfs: Connection timed out
$

[Poseidon is my server at 192.168.1.6, the client is at 192.168.1.2]

Server disk to be exported is a ~9TB raid array with XFS.

I'm using nfs3 with ACL and no idmapd; nfs4+ is not compiled into kernel 
(neither on client nor on server); Why it is trying nfs4 first as seen in the 
log above I don't know. nfs-utils has been compiled with USE=-nfsv4

Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9 
installed. As both clients and server are not accessable from outside, no 
firewalls are installed.

What I checked:
/etc/exports:
/datadisk               192.168.1.0/24(rw,async,subtree_check)                                                                
        
portmapper, nfs-services are running normal, as far I can see.

Does anyone have any suggestion?

Thanks,
        Alex

Re: [gentoo-user] mount.nfs stale nfs handle

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 666 bytes --]

On Sun, 29 Jun 2014 21:34:07 +0200, Alexander Puchmayr wrote:

> After upgrading my server to latest stable release of gentoo, none of
> my clients is able to mount any nfs share from the server anymore.

[snip]

> Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9 
> installed. As both clients and server are not accessable from outside,
> no firewalls are installed.

That's not the latest nfs-utils in stable, it is 1.2.9-r3.

Are you using openrc or systemd? There was a problem with some
systemd service files in recent nfs-utils releases, fixed now.


-- 
Neil Bothwick

I'm not anti-social, I'm just not user friendly

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: [gentoo-user] mount.nfs stale nfs handle

[Alexander Puchmayr]

Am Sonntag, 29. Juni 2014, 20:41:55 schrieb Neil Bothwick:
> On Sun, 29 Jun 2014 21:34:07 +0200, Alexander Puchmayr wrote:
> > After upgrading my server to latest stable release of gentoo, none of
> > my clients is able to mount any nfs share from the server anymore.
> 
> [snip]
> 
> > Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9
> > installed. As both clients and server are not accessable from outside,
> > no firewalls are installed.
> 
> That's not the latest nfs-utils in stable, it is 1.2.9-r3.
> 

This morning it was masked; OK, emerge --sync && emerge nfs-utils.
After restarting all nfs relevant services it is still the same :-(

> Are you using openrc or systemd? There was a problem with some
> systemd service files in recent nfs-utils releases, fixed now.

I'm using openrc.

Alex

Re: [gentoo-user] mount.nfs stale nfs handle

[Joost Roeleveld]

On Sunday 29 June 2014 22:48:32 Alexander Puchmayr wrote:
> Am Sonntag, 29. Juni 2014, 20:41:55 schrieb Neil Bothwick:
> > On Sun, 29 Jun 2014 21:34:07 +0200, Alexander Puchmayr wrote:
> > > After upgrading my server to latest stable release of gentoo, none of
> > > my clients is able to mount any nfs share from the server anymore.
> > 
> > [snip]
> > 
> > > Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9
> > > installed. As both clients and server are not accessable from outside,
> > > no firewalls are installed.
> > 
> > That's not the latest nfs-utils in stable, it is 1.2.9-r3.
> 
> This morning it was masked; OK, emerge --sync && emerge nfs-utils.
> After restarting all nfs relevant services it is still the same :-(

Just to confirm, did you update nfs-utils on both systems?

--
Joost

Re: [gentoo-user] mount.nfs stale nfs handle

[Joost Roeleveld]

On Sunday 29 June 2014 21:34:07 Alexander Puchmayr wrote:
> Hi there,
> 
> After upgrading my server to latest stable release of gentoo, none of my
> clients is able to mount any nfs share from the server anymore.
> 
> Symptoms:
> $ mount -v -t nfs poseidon:/datadisk/ /mnt/gentoo/
> mount.nfs: timeout set for Sun Jun 29 19:33:40 2014
> mount.nfs: trying text-based options
> 'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
> mount.nfs: mount(2): Protocol not supported
> mount.nfs: trying text-based options 'addr=192.168.1.6'
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=17
> mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
> mount.nfs: mount(2): Stale NFS file handle
> mount.nfs: trying text-based options
> 'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
> mount.nfs: mount(2): Protocol not supported
> mount.nfs: trying text-based options 'addr=192.168.1.6'
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=17
> mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
> mount.nfs: mount(2): Stale NFS file handle
> [...]
> mount.nfs: Connection timed out
> $
> 
> [Poseidon is my server at 192.168.1.6, the client is at 192.168.1.2]
> 
> Server disk to be exported is a ~9TB raid array with XFS.
> 
> I'm using nfs3 with ACL and no idmapd; nfs4+ is not compiled into kernel
> (neither on client nor on server); Why it is trying nfs4 first as seen in
> the log above I don't know. nfs-utils has been compiled with USE=-nfsv4
> 
> Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9
> installed. As both clients and server are not accessable from outside, no
> firewalls are installed.
> 
> What I checked:
> /etc/exports:
> /datadisk               192.168.1.0/24(rw,async,subtree_check)
> 
> portmapper, nfs-services are running normal, as far I can see.
> 
> Does anyone have any suggestion?

I have this occasionally due to the backup system I am using:
- stop the nfs export
- umount the filesystem
- take LVM snapshot
- remound filesystem
- re-enable the nfs export

When that happens, I run the following on the server:
# exportfs -au && sleep 1 && mount -a && sleep 1 && exportfs -r

The sleeps are necessary, without them, it doesn't always work.

--
Joost

Re: [gentoo-user] mount.nfs stale nfs handle

[Alexander Puchmayr]

Yes. Both client and server have the actual version.

Alex

Gesendet mit AquaMail für Android
http://www.aqua-mail.com


Am 30. Juni 2014 09:30:01 schrieb Joost Roeleveld <joost@antarean.org>:

> On Sunday 29 June 2014 22:48:32 Alexander Puchmayr wrote:
> > Am Sonntag, 29. Juni 2014, 20:41:55 schrieb Neil Bothwick:
> > > On Sun, 29 Jun 2014 21:34:07 +0200, Alexander Puchmayr wrote:
> > > > After upgrading my server to latest stable release of gentoo, none of
> > > > my clients is able to mount any nfs share from the server anymore.
> > >
> > > [snip]
> > >
> > > > Server has kernel version 3.12.21-gentoo-r1and net-fs/nfs-utils-1.2.9
> > > > installed. As both clients and server are not accessable from outside,
> > > > no firewalls are installed.
> > >
> > > That's not the latest nfs-utils in stable, it is 1.2.9-r3.
> >
> > This morning it was masked; OK, emerge --sync && emerge nfs-utils.
> > After restarting all nfs relevant services it is still the same :-(
>
> Just to confirm, did you update nfs-utils on both systems?
>
> --
> Joost
>
>
> !DSPAM:506,53b10f55576434764113543!
>
>

Re: [gentoo-user] mount.nfs stale nfs handle [SOLVED]

[Alexander Puchmayr]

Am Sonntag, 29. Juni 2014, 21:34:07 schrieb Alexander Puchmayr:
> Hi there,
> 
> After upgrading my server to latest stable release of gentoo, none of my
> clients is able to mount any nfs share from the server anymore.
> 
> Symptoms:
> $ mount -v -t nfs poseidon:/datadisk/ /mnt/gentoo/
> mount.nfs: timeout set for Sun Jun 29 19:33:40 2014
> mount.nfs: trying text-based options
> 'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
> mount.nfs: mount(2): Protocol not supported
> mount.nfs: trying text-based options 'addr=192.168.1.6'
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=17
> mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
> mount.nfs: mount(2): Stale NFS file handle
> mount.nfs: trying text-based options
> 'vers=4,addr=192.168.1.6,clientaddr=192.168.1.2'
> mount.nfs: mount(2): Protocol not supported
> mount.nfs: trying text-based options 'addr=192.168.1.6'
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying 192.168.1.6 prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=17
> mount.nfs: trying 192.168.1.6 prog 100005 vers 3 prot UDP port 60058
> mount.nfs: mount(2): Stale NFS file handle
> [...]
> mount.nfs: Connection timed out
> $
> 
[...]

The important hint I finally found after searching for hours:
> http://www.mmacleod.ca/blog/2014/02/nfs-exports-and-xfss-inode64-mount-option/
Thanks to the guy who wrote it!

The key is that
* I'm exporting more than one different (sub-)directories on the same filesystem
* For some reason it gets confused with the uuid/fsid; Can't find the page 
where I found that anymore :-(
* the actual device the exported filesystem resides on is /dev/md127 and not 
/dev/md0 as I originally wanted. Since I did not regard the number, I ignored 
this fact, but I also found hints that this might cause the problem of the 
previous point. 

The suggested solution was to add a *unique* fsid=xx entry to the exports file 
for each directory exported, so that it looks like:

/etc/exports:
/datadisk/music	192.168.1.0/24(rw,no_subtree_check,fsid=1)
/datadisk/video		192.168.1.0/24(rw,no_subtree_check,fsid=2)
/datadisk/backup	192.168.1.0/24(ro,no_subtree_check,fsid=3)
...

If the same fsid is used more than once then the first directory with this fsid 
will be mounted! So using *different* fsids for each exported directory is 
essential!!!

So, my NFS works now as it should! 

Thanks to all who spent a thought on it!

Alex