NFS4+KRB5 trouble in Kernel >3.8

NFS4+KRB5 trouble in Kernel >3.8

[Sven Geggus]

Hello,

seems I have been bitten again by the same Problem as back in November
(http://comments.gmane.org/gmane.linux.nfs/52916).

Well, at least the symptoms are the same. The problem is almost
certainly another one.

In recent kernels (e.g. 3.9.4) again my mount seems to hang infinitely:

mount -t nfs4 -o sec=krb5 testsrv:/storage /mnt/tmp 

This time I immediately tried to bisect this, but what happens is that half
the way down bisecting the symptoms change. The mount does not hang anymore,
but I get "access denied" now:

 ~/ # mount -t nfs4 -o sec=krb5 testsrv:/storage /mnt/tmp
mount.nfs4: access denied by server while mounting testsrv:/storage

So how to continue bisecting? Should I mark this as good or bad?

Here is the output of "git bisect log" so far:

git bisect start
# good: [dbf932a9b316d5b29b3e220e5a30e7a165ad2992] Linux 3.8.13
git bisect good dbf932a9b316d5b29b3e220e5a30e7a165ad2992
# bad: [0bfd8ffeff9dda08c69381d65c72e0aa58706ef6] Linux 3.9.4
git bisect bad 0bfd8ffeff9dda08c69381d65c72e0aa58706ef6
# good: [19f949f52599ba7c3f67a5897ac6be14bfcb1200] Linux 3.8
git bisect good 19f949f52599ba7c3f67a5897ac6be14bfcb1200
# good: [ef53d16cded7f89b3843b7a96970dab897843ea5] ksm: allocate roots when needed
git bisect good ef53d16cded7f89b3843b7a96970dab897843ea5

I did not mark the current state with either good or bad yet.

Regards

Sven

-- 
Software patents are the software project equivalent of land mines: Each
design decision carries a risk of stepping on a patent, which can destroy
your project. (Richard M. Stallman)
/me is giggls@ircnet, http://sven.gegg.us/ on the Web

Re: NFS4+KRB5 trouble in Kernel >3.8

[Chuck Lever]

On Jun 6, 2013, at 9:24 AM, Sven Geggus <lists@fuchsschwanzdomain.de> wrote:

> Hello,
> 
> seems I have been bitten again by the same Problem as back in November
> (http://comments.gmane.org/gmane.linux.nfs/52916).
> 
> Well, at least the symptoms are the same. The problem is almost
> certainly another one.
> 
> In recent kernels (e.g. 3.9.4) again my mount seems to hang infinitely:
> 
> mount -t nfs4 -o sec=krb5 testsrv:/storage /mnt/tmp 
> 
> This time I immediately tried to bisect this,

Perhaps you could try enabling debugging first to see where the hang occurs.

  $ sudo rpcdebug -m nfs -s all

then add "-vvv" on the rpc.gssd command line.  Try your mount again, and look in /var/log/messages.


> but what happens is that half
> the way down bisecting the symptoms change. The mount does not hang anymore,
> but I get "access denied" now:
> 
> ~/ # mount -t nfs4 -o sec=krb5 testsrv:/storage /mnt/tmp
> mount.nfs4: access denied by server while mounting testsrv:/storage
> 
> So how to continue bisecting? Should I mark this as good or bad?
> 
> Here is the output of "git bisect log" so far:
> 
> git bisect start
> # good: [dbf932a9b316d5b29b3e220e5a30e7a165ad2992] Linux 3.8.13
> git bisect good dbf932a9b316d5b29b3e220e5a30e7a165ad2992
> # bad: [0bfd8ffeff9dda08c69381d65c72e0aa58706ef6] Linux 3.9.4
> git bisect bad 0bfd8ffeff9dda08c69381d65c72e0aa58706ef6
> # good: [19f949f52599ba7c3f67a5897ac6be14bfcb1200] Linux 3.8
> git bisect good 19f949f52599ba7c3f67a5897ac6be14bfcb1200
> # good: [ef53d16cded7f89b3843b7a96970dab897843ea5] ksm: allocate roots when needed
> git bisect good ef53d16cded7f89b3843b7a96970dab897843ea5
> 
> I did not mark the current state with either good or bad yet.
> 
> Regards
> 
> Sven
> 
> -- 
> Software patents are the software project equivalent of land mines: Each
> design decision carries a risk of stepping on a patent, which can destroy
> your project. (Richard M. Stallman)
> /me is giggls@ircnet, http://sven.gegg.us/ on the Web
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Chuck Lever
chuck[dot]lever[at]oracle[dot]com

Re: NFS4+KRB5 trouble in Kernel >3.8

[Sven Geggus]

Chuck Lever <chuck.lever@oracle.com> wrote:

> Perhaps you could try enabling debugging first to see where the hang occurs.
> 
>  $ sudo rpcdebug -m nfs -s all
> 
> then add "-vvv" on the rpc.gssd command line.  Try your mount again, and look in /var/log/messages.

Client or server side or both?

As you have been talking about rpc.gssd I assume client side, right?

Sven

-- 
Software patents are the software project equivalent of land mines: Each
design decision carries a risk of stepping on a patent, which can destroy
your project. (Richard M. Stallman)
/me is giggls@ircnet, http://sven.gegg.us/ on the Web

Re: NFS4+KRB5 trouble in Kernel >3.8

[Chuck Lever]

On Jun 7, 2013, at 6:19 AM, Sven Geggus <lists@fuchsschwanzdomain.de> wrote:

> Chuck Lever <chuck.lever@oracle.com> wrote:
> 
>> Perhaps you could try enabling debugging first to see where the hang occurs.
>> 
>> $ sudo rpcdebug -m nfs -s all
>> 
>> then add "-vvv" on the rpc.gssd command line.  Try your mount again, and look in /var/log/messages.
> 
> Client or server side or both?
> 
> As you have been talking about rpc.gssd I assume client side, right?

Your complaint was about mount problems, so I assumed client side, and did not visit the link you provided.

>From memory, the server equivalent is "-vv" or "-vvv" on the rpc.svcgssd command line, and

  $ sudo rpcdebug -m nfsd -s all

You might also enable debugging on mountd, see rpc.mountd(8).

-- 
Chuck Lever
chuck[dot]lever[at]oracle[dot]com

Re: NFS4+KRB5 trouble in Kernel >3.8

[Sven Geggus]

Chuck Lever <chuck.lever@oracle.com> wrote:

> Your complaint was about mount problems, so I assumed client side,
> and did not visit the link you provided.

My Problem is almost certainly a server Problem. I will see what I
can find with tcpdump/wireshark on monday.

Sven

-- 
"Dynamische IP-Nummern sind Security-Homöopathie."
						(Kristian Köhntopp)

/me is giggls@ircnet, http://sven.gegg.us/ on the Web