Linux NFS development
 help / color / mirror / Atom feed
From: "François Valenduc" <francois.valenduc@skynet.be>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Kevin Coffman <kwc@umich.edu>, linux-nfs@vger.kernel.org
Subject: Re: nfs and kerberos authentification problem.
Date: Thu, 04 Sep 2008 19:31:11 +0200	[thread overview]
Message-ID: <48C01B5F.3060808@skynet.be> (raw)
In-Reply-To: <20080904165645.GG4536@fieldses.org>

J. Bruce Fields a =C3=A9crit :
> On Thu, Sep 04, 2008 at 06:45:03PM +0200, Fran=C3=A7ois Valenduc wrot=
e:
>  =20
>> Kevin Coffman a =C3=A9crit :
>>    =20
>>> Hello Fran=C3=A7ois,
>>> First, you should not need to limit the encryption types in
>>> /etc/krb5.conf as you have done.  None of the following lines are
>>> necessary in either the client or server's /etc/krb5.conf file.
>>> (Leaving them in will probably lead to headaches with other Kerbero=
s
>>> applications in the future.)
>>>
>>>   default_tkt_enctypes =3D aes256-cts-hmac-sha1-96 des-cbc-crc
>>>   default_tgs_enctypes =3D aes256-cts-hmac-sha1-96 des-cbc-crc
>>>   permitted_enctypes =3D aes256-cts-hmac-sha1-96 rc4-hmac des-cbc-c=
rc
>>> des3-hmac-sha1
>>>
>>> You said that you limited the client's keytab to des-cbc-crc.  It
>>> appears you have done the same for the server's keytab since the
>>> ticket and session key the client gets are des-cbc-crc.
>>>
>>>  =20
>>>      =20
>>>> Sep  3 19:36:22 pc-francois krb5kdc[9787]: TGS_REQ (2 etypes {18 1=
})
>>>> 192.168.1.3: ISSUE: authtime 1220463382, etypes {rep=3D18 tkt=3D1 =
ses=3D1},
>>>> nfs/ordi-francois.homenetwork.net-wmZDWbG+120CDknkFGB/9A@public.gmane.org for
>>>> nfs/pc-francois.homenetwork.net-wmZDWbG+120CDknkFGB/9A@public.gmane.org
>>>>    =20
>>>>        =20
>>> It looks like the client is successfully authenticating as
>>> "nfs/ordi-francois.homenetwork.net-wmZDWbG+120CDknkFGB/9A@public.gmane.org".
>>>
>>>  =20
>>>      =20
>>>> Sep  3 19:36:22 pc-francois rpc.svcgssd[7008]: sname =3D
>>>> nfs/ordi-francois.homenetwork.net-wmZDWbG+120CDknkFGB/9A@public.gmane.org
>>>>    =20
>>>>        =20
>>> However, mapping that gss_auth_name to a local ID is failing, and i=
s
>>> being mapped to uid/gid of "-1 -1"
>>> .
>>>
>>>  =20
>>>      =20
>>>> Sep  3 19:36:22 pc-francois rpc.svcgssd[7008]: \x01000000 21474836=
47 -1 -1 0 krb5
>>>> \x000000000000[...]80b98
>>>>    =20
>>>>        =20
>>> The "-1" should be interpreted in the kernel as nfsnobody.  What ar=
e
>>> the permissions on the exported filesystem?
>>>
>>> K.C.
>>>
>>>  =20
>>>      =20
>> So, is it normat that gss map the local uid/gid to -1 -1 ? If not, w=
hat =20
>> should I change ?
>> The folder I try to export is configured like this:
>>
>> /home/francois ordi-francois(rw,root_squash,no_subtree_check)
>>    =20
>
> Assyming you're using nfs-utils 1.1.1 or later, add "sec=3Dkrb5" to t=
he
> export options.  (Or "sec=3Dkrb5:krb5i:krb5p" if you also want to all=
ow
> integrity and privacy; and "sec=3Dsys:krb5:krb5i:krb5p" if you also w=
ant
> to allow auth_sys.)
>
> --b.
>
>  =20
I had indeed forgot to add sec=3Dkrb5 to the export options. But even i=
f I=20
add it, it doesn't change anything. Is it really possible to use krb5=20
authentification with nfs ? I have read a lot of howto and follow the=20
instructions and it never succeeds...

=46ran=C3=A7ois

  reply	other threads:[~2008-09-04 17:31 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-09-03 18:19 nfs and kerberos authentification problem François Valenduc
2008-09-03 20:12 ` Kevin Coffman
     [not found]   ` <4d569c330809031312p3515f4d8id9cbec94d871e058-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2008-09-04 16:45     ` François Valenduc
2008-09-04 16:56       ` J. Bruce Fields
2008-09-04 17:31         ` François Valenduc [this message]
2008-09-04 17:33           ` J. Bruce Fields
2008-09-04 17:41             ` François Valenduc
2008-09-04 17:49               ` J. Bruce Fields
2008-09-04 17:58                 ` François Valenduc
2008-09-04 18:39                   ` J. Bruce Fields
2008-09-04 18:53                     ` François Valenduc
2008-09-04 18:59                       ` J. Bruce Fields
2008-09-04 19:31                         ` Kevin Coffman
     [not found]                           ` <4d569c330809041231wcbddde8w419968280de9e39a-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2008-09-04 19:38                             ` François Valenduc
2008-09-04 19:40                               ` J. Bruce Fields
2008-09-04 19:41                                 ` J. Bruce Fields
2008-09-04 19:56                                   ` François Valenduc
2008-09-05 18:36                                 ` François Valenduc
2008-09-05 18:57                                   ` François Valenduc
2008-09-05 21:26                                     ` J. Bruce Fields
2008-09-05 21:23                                   ` J. Bruce Fields

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=48C01B5F.3060808@skynet.be \
    --to=francois.valenduc@skynet.be \
    --cc=bfields@fieldses.org \
    --cc=kwc@umich.edu \
    --cc=linux-nfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox