Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
Hi,

I've launched thinclients on Skole, I have splashy and as usual it crached
at one moment, but then I don't have verbose and if I wait I have this
message :

[215.042458] Kernel panic - not syncing : Attempt to kill init !

What does it mean on a thinclient ?
Does anybody had this ?

I have 3 switchs between tjener and this thinclient.

-- 
Bien cordialement/ Cheers

Jean-Charles Siegel


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
So I have to quit/kill splashy so I can have all verbose ?

Because when splashy stops, I don't have any text except : [215.042458]
Kernel panic - not syncing : Attempt to kill init !

it's the only thing I have on the screen.

2011/1/5 Ronny Aasen 

> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> On 05. jan. 2011 11:23, Jean-Charles Skolelinux wrote:
> > Hi,
> >
> > I've launched thinclients on Skole, I have splashy and as usual it
> > crached at one moment, but then I don't have verbose and if I wait I
> > have this message :
> >
> > [215.042458] Kernel panic - not syncing : Attempt to kill init !
> >
> > What does it mean on a thinclient ?
> > Does anybody had this ?
> >
> > I have 3 switchs between tjener and this thinclient.
> >
> > --
> > Bien cordialement/ Cheers
> >
> > Jean-Charles Siegel
> >
>
>
> that is usualy the last line, saying init died. the actual errors are
> further up in the text.
>
> you want to look at earlier error message. usualy something about why it
> could not run init.
>
> Ronny Aasen
>
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAk0kSwUACgkQdjPGjuyRrjou/ACg+JAIxtcsyvyI1Rz97U2BBTcf
> 7XkAoJgHGLq3OR96JiTEdDM+P6AfTPCW
> =9ZU9
> -END PGP SIGNATURE-
>



-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
And I have to wait a few minutes before having kernel panic message.

2011/1/5 Jean-Charles Skolelinux 

> So I have to quit/kill splashy so I can have all verbose ?
>
> Because when splashy stops, I don't have any text except : [215.042458]
> Kernel panic - not syncing : Attempt to kill init !
>
> it's the only thing I have on the screen.
>
> 2011/1/5 Ronny Aasen 
>
> -BEGIN PGP SIGNED MESSAGE-
>> Hash: SHA1
>>
>> On 05. jan. 2011 11:23, Jean-Charles Skolelinux wrote:
>> > Hi,
>> >
>> > I've launched thinclients on Skole, I have splashy and as usual it
>> > crached at one moment, but then I don't have verbose and if I wait I
>> > have this message :
>> >
>> > [215.042458] Kernel panic - not syncing : Attempt to kill init !
>> >
>> > What does it mean on a thinclient ?
>> > Does anybody had this ?
>> >
>> > I have 3 switchs between tjener and this thinclient.
>> >
>> > --
>> > Bien cordialement/ Cheers
>> >
>> > Jean-Charles Siegel
>> >
>>
>>
>> that is usualy the last line, saying init died. the actual errors are
>> further up in the text.
>>
>> you want to look at earlier error message. usualy something about why it
>> could not run init.
>>
>> Ronny Aasen
>>
>> -BEGIN PGP SIGNATURE-
>> Version: GnuPG v1.4.10 (GNU/Linux)
>> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>>
>> iEYEARECAAYFAk0kSwUACgkQdjPGjuyRrjou/ACg+JAIxtcsyvyI1Rz97U2BBTcf
>> 7XkAoJgHGLq3OR96JiTEdDM+P6AfTPCW
>> =9ZU9
>> -END PGP SIGNATURE-
>>
>
>
>
> --
> Bien cordialement/ Cheers
>
> Jean-Charles Siegel
> French Skolelinux Team
>
>


-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 11.23.34 skrev Jean-Charles Skolelinux :
> [215.042458] Kernel panic - not syncing : Attempt to kill init !
> 
> What does it mean on a thinclient ?
> Does anybody had this ?

If this is a new client, that you never had in production before, then 
this "normally" means that you have no support for the networkcard in 
your initrd.

If this client used to work before, but now suddenly it says this, then 
that is a different situation.

Which is it?
-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


Re: Kernel Panic

2011-01-05 Thread Ronny Aasen
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On 05. jan. 2011 11:23, Jean-Charles Skolelinux wrote:
> Hi,
> 
> I've launched thinclients on Skole, I have splashy and as usual it
> crached at one moment, but then I don't have verbose and if I wait I
> have this message :
> 
> [215.042458] Kernel panic - not syncing : Attempt to kill init !
> 
> What does it mean on a thinclient ?
> Does anybody had this ?
> 
> I have 3 switchs between tjener and this thinclient.
> 
> -- 
> Bien cordialement/ Cheers
> 
> Jean-Charles Siegel
> 


that is usualy the last line, saying init died. the actual errors are
further up in the text.

you want to look at earlier error message. usualy something about why it
could not run init.

Ronny Aasen

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk0kSwUACgkQdjPGjuyRrjou/ACg+JAIxtcsyvyI1Rz97U2BBTcf
7XkAoJgHGLq3OR96JiTEdDM+P6AfTPCW
=9ZU9
-END PGP SIGNATURE-


-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4d244b09.6060...@skolelinux.no



Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
This client was not in production before. BUT this client comes from a serie
of computer (Compaq Evo), the network card is a 3com.

Is this error could come because the MAC address of my thinclient is the
same as an other thinclient ?

2011/1/5 Klaus Ade Johnstad 

>  Onsdag 5. januar 2011 11.23.34 skrev Jean-Charles Skolelinux :
> > [215.042458] Kernel panic - not syncing : Attempt to kill init !
> >
> > What does it mean on a thinclient ?
> > Does anybody had this ?
>
> If this is a new client, that you never had in production before, then
> this "normally" means that you have no support for the networkcard in
> your initrd.
>
> If this client used to work before, but now suddenly it says this, then
> that is a different situation.
>
> Which is it?
> --
> Klaus Ade
> kl...@bzz.no
> 67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D
>



-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Rosario Crispo

Hi,

I had the same problem with amd cpu on asus pc.


Try different ram it sounds like the ram is either not compatible with 
the motherboard or bad or set into the bios the correct speed ram (I 
solved with last option).



Ross



Il 2011/01/05 11:49, Klaus Ade Johnstad ha scritto:


  Onsdag 5. januar 2011 11.23.34 skrev Jean-Charles Skolelinux :

[215.042458] Kernel panic - not syncing : Attempt to kill init !

What does it mean on a thinclient ?
Does anybody had this ?

If this is a new client, that you never had in production before, then
this "normally" means that you have no support for the networkcard in
your initrd.

If this client used to work before, but now suddenly it says this, then
that is a different situation.

Which is it?



--
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4d245121.5050...@gmail.com



Re: Kernel Panic

2011-01-05 Thread Rosario Crispo

Hi,

I had the same problem with amd cpu on asus pc.


Try different ram it sounds like the ram is either not compatible with 
the motherboard or bad or set into the bios the correct speed ram (I 
solved with last option).



cheers


Ross



Il 2011/01/05 11:49, Klaus Ade Johnstad ha scritto:


  Onsdag 5. januar 2011 11.23.34 skrev Jean-Charles Skolelinux :

[215.042458] Kernel panic - not syncing : Attempt to kill init !

What does it mean on a thinclient ?
Does anybody had this ?

If this is a new client, that you never had in production before, then
this "normally" means that you have no support for the networkcard in
your initrd.

If this client used to work before, but now suddenly it says this, then
that is a different situation.

Which is it?



--
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4d24515f.3000...@gmail.com



Re: Kernel Panic

2011-01-05 Thread Wolfgang Schweer
Hello Jean-Charles,

On Mi, 05 Jan 2011, Jean-Charles Siegel wrote:

> This client was not in production before. BUT this client comes from
> a serie of computer (Compaq Evo), the network card is a 3com.
 
Is it 3cR990? If you're testing debian-edu squeeze the failure is due 
to missing package firmware-linux-nonfree in the ltsp-chroot.
(as a consequence the initrd doesn't contain the binary part of the 
modul typhoon.)

Wolfgang


--
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20110105111830.ga6...@schweer-online.local



Re: Kernel Panic

2011-01-05 Thread Jonas Smedegaard

On Wed, Jan 05, 2011 at 11:23:34AM +0100, Jean-Charles Skolelinux wrote:
I've launched thinclients on Skole, I have splashy and as usual it 
crached at one moment, but then I don't have verbose and if I wait I 
have this message :


[215.042458] Kernel panic - not syncing : Attempt to kill init !

What does it mean on a thinclient ?


First off, you need to disable splashy so as to not hide the boot 
messages.  They are crucial in figuring out what happened right before 
the panic.



 - Jonas

--
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: Digital signature


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
I don't know how disable splashy :(
I'm looking for a tutorial but I don't find...

2011/1/5 Jonas Smedegaard 

> On Wed, Jan 05, 2011 at 11:23:34AM +0100, Jean-Charles Skolelinux wrote:
>
>> I've launched thinclients on Skole, I have splashy and as usual it crached
>> at one moment, but then I don't have verbose and if I wait I have this
>> message :
>>
>> [215.042458] Kernel panic - not syncing : Attempt to kill init !
>>
>> What does it mean on a thinclient ?
>>
>
> First off, you need to disable splashy so as to not hide the boot messages.
>  They are crucial in figuring out what happened right before the panic.
>
>
>  - Jonas
>
> --
>  * Jonas Smedegaard - idealist & Internet-arkitekt
>  * Tlf.: +45 40843136  Website: http://dr.jones.dk/
>
>  [x] quote me freely  [ ] ask before reusing  [ ] keep private
>
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.10 (GNU/Linux)
>
> iQIcBAEBCgAGBQJNJFmdAAoJECx8MUbBoAEh/4UQAJMyiZ+99vZd+F4fJlcai6BY
> Yxrghkb72JmaGC7Sv0otUAEThGrmIvofy5O8M01eveYq/pjRqpXKJP+vGxqILdp0
> QcBTriwUJJZu6OT2+YhkYHOv80W64esmEKvJBl9FM6SSDPCjSSBFt17P3svADxGQ
> AggIUoH64e4zMdgfC78QhSFubAKp2YoM2Pfxcoo/6nZ5lZfL9Ync2HCDxg3Ll7UW
> k4WHOLHAawfOknKABHvH/QjdOVPU1GJCI3Dk4IGzV5TN3NcSPtVwBJcptFMxDL18
> fcu87Euc/N6IlLGLVu9/I2apI8OFwhiHKWJG1KHT2YuknDqTaWUZNZXu5zGDcNBz
> rEGusp0OfHfJjLkZg7J7aem9Bg/5HS4b+OPN+FRtwOQMmueEnEbMY9ju+VgsYfEa
> b66RqwqImTdul9A8o9bNOUpzicA/Fk3ITgeGYo9Ldh18ao3mA9MdrrtASqPHBChC
> MGaz6BGpIUElghaDv3bPiYUe7p6TLrW2f59p7ayH5Y0xvx1gTL/0krPDVAVsDKPp
> avQreppfplsdlAqOUgUtmjZ3bQYXlzi9yzyYiBtO/u3zPj0vUJFyGM6xUJNkjRV1
> v3MQJIGwvVasH/Od/1o9A7NaWXI9ofiCZ/4R1Oe+M1zNAgIFlCoPgHqiZ/hkFkZJ
> COk1jQxcdkAxuKIQQG4P
> =ZoQX
> -END PGP SIGNATURE-
>
>


-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Petter Reinholdtsen
[Jean-Charles Skolelinux]
> I don't know how disable splashy :(
> I'm looking for a tutorial but I don't find...

Try booting with the kernel argument 'nosplash'.

Happy hacking,
-- 
Petter Reinholdtsen


-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20110105122556.gq31...@login2.uio.no



Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 13.22.43 skrev Jean-Charles Skolelinux :
> I don't know how disable splashy :(
> I'm looking for a tutorial but I don't find...

Look in the file "default", either in 
/var/lib/tftpboot/pxelinux.cfg/default or in 
/var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default

(If you can't see it there, look for it, with something like 
"locate default|grep pxelinux.cfg")

There just add/edit the word "nosplash" add the end of the line that 
starts with "append". 

If you only have one line in default-file, then just add "nosplash" at 
the end of that line.

Make sure that you do not have the word "splash" there.

 
-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 13.43.15 skrev Jean-Charles Skolelinux :
> I have change /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
> 
> I had this : DEFAULT vmlinuz ro initrd=initrd.img boot=nfs quiet
> 
> I add nosplash.
> 
> The message is : nfsmount : need a path
> 
> But it's only on this thinclient, on other it's starting.

Didn't you ask for howto about changing ip-adresse the other day? Did 
you try to change the ip-adresse? If yes, then probably somewhere you 
made a mistake. No valid root-path explains a lot.
-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
No for now, I don't change Ip-address and I won't do that on a production
server before I test it

2011/1/5 Klaus Ade Johnstad 

>  Onsdag 5. januar 2011 13.43.15 skrev Jean-Charles Skolelinux :
> > I have change /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
> >
> > I had this : DEFAULT vmlinuz ro initrd=initrd.img boot=nfs quiet
> >
> > I add nosplash.
> >
> > The message is : nfsmount : need a path
> >
> > But it's only on this thinclient, on other it's starting.
>
> Didn't you ask for howto about changing ip-adresse the other day? Did
> you try to change the ip-adresse? If yes, then probably somewhere you
> made a mistake. No valid root-path explains a lot.
> --
> Klaus Ade
> kl...@bzz.no
> 67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D
>



-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 14.01.47 skrev Jean-Charles Skolelinux :
> No for now, I don't change Ip-address and I won't do that on a
> production server before I test it

Maybe you can give us the few lines from syslog when the thinclient 
boots?
-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
I change thinclient. I'm on a eVectra from HP

I have :

IP config : eth0 hardware address 00:00:02:02:67:0a mtu 1500 DHCP RARP

and then after a few minutes I obtain an address from 10.0.2.2 and not from
192.168.0.254.

Just after that I have the message : nfsmount : need a path

but other thinclients like this one start on the same network

2011/1/5 Klaus Ade Johnstad 

>  Onsdag 5. januar 2011 14.01.47 skrev Jean-Charles Skolelinux :
> > No for now, I don't change Ip-address and I won't do that on a
> > production server before I test it
>
> Maybe you can give us the few lines from syslog when the thinclient
> boots?
> --
> Klaus Ade
> kl...@bzz.no
> 67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D
>



-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 14.34.05 skrev Jean-Charles Skolelinux :
> and then after a few minutes I obtain an address from 10.0.2.2 and
> not from 192.168.0.254.

Then either you have misconfiguration in your network (eth0 and eth1 on 
the server is somehow connected), or you have your thinclients on the 
"wrong" network (strange, I thought they should be able to boot on both 
10-network aswell as on 192-network).


-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


Re: Kernel Panic

2011-01-05 Thread Jean-Charles Skolelinux
I beg your pardon, I've wasted your time.

The mistake is human, but skole is perfect.

My mistake was to use a routeur as a switch and I forgot to shutdown the
dhcp server on the routeur, so there were conflicts between the dhcp of
tjener and the dhcp of the routeur.

I thank all of you very much for your help and your time.

2011/1/5 Klaus Ade Johnstad 

>  Onsdag 5. januar 2011 14.34.05 skrev Jean-Charles Skolelinux :
> > and then after a few minutes I obtain an address from 10.0.2.2 and
> > not from 192.168.0.254.
>
> Then either you have misconfiguration in your network (eth0 and eth1 on
> the server is somehow connected), or you have your thinclients on the
> "wrong" network (strange, I thought they should be able to boot on both
> 10-network aswell as on 192-network).
>
>
> --
> Klaus Ade
> kl...@bzz.no
> 67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D
>



-- 
Bien cordialement/ Cheers

Jean-Charles Siegel
French Skolelinux Team


Re: Kernel Panic

2011-01-05 Thread Klaus Ade Johnstad
 Onsdag 5. januar 2011 16.05.14 skrev Jean-Charles Skolelinux :
> I beg your pardon, I've wasted your time.

False.

> The mistake is human, but skole is perfect.

True.

:-)

-- 
Klaus Ade
kl...@bzz.no
67E61D18B2C44F8A3DA35C6D849F9F5F 26FA477D


signature.asc
Description: This is a digitally signed message part.


NFS4 and Kerberos: A-records for same IP inflate the need for service principals

2011-01-05 Thread Andreas B. Mundt
Hi all,

the last days I found a little time to have a look into the issue of
using NFSv4 (and perhaps Kerberos) to mount the home directories.

I first configured NFS4 to export the home directories. After that I
tried kerberos authentication. However, I observed that it works only
in some cases, in most of the attempts to mount the share a missing
principal of the form nfs/x...@intern was reported, where XXX is one of
the hostnames (and not tjener.intern) reported by this command: 

r...@tjener:~# host 10.0.2.2
2.2.0.10.in-addr.arpa domain name pointer tjener.intern.
2.2.0.10.in-addr.arpa domain name pointer kerberos.intern.
2.2.0.10.in-addr.arpa domain name pointer ldap.intern.
2.2.0.10.in-addr.arpa domain name pointer domain.intern.
2.2.0.10.in-addr.arpa domain name pointer postoffice.intern.
2.2.0.10.in-addr.arpa domain name pointer syslog.intern.

If I understand things correctly, mounting the share with 

mount -t nfs4 -o sec=krb5 tjener.intern:/ /skole/tjener/

converts tjener.intern into an IP adress and that address back to the
(full qualified) hostname. So only if by chance tjener.intern is used
for the lookup, the (existing) nfs/tjener.int...@intern principal is
used and things work as they should. If another hostname is used,
things fail because there is no corresponding service principal.

I tried to find the reason for these corresponding A-records, they
have been changed in commit 71704.
(http://svn.debian.org/wsvn/debian-edu/trunk/src/debian-edu-config/ldap-bootstrap/?rev=71704&sc=1>)
 

I am not an expert regarding that stuff and I don't know if there are
other ways to achieve the desired. However, it looks as with the
current setup we need service principals for all host aliases.

Best regards,

 Andi
 




-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20110105175842.ga4...@flashgordon



Re: NFS4 and Kerberos: A-records for same IP inflate the need for service principals

2011-01-05 Thread Petter Reinholdtsen
[Andreas B. Mundt]
> I tried to find the reason for these corresponding A-records,

There are two aspects coming together to cause this effect.  We use
strict mode in powerdns, allowing shared A/PTR entries in LDAP, and
the fact that SRV and MX records need to point to A records.  As our
design allow for scaling by moving individual services out by changing
DNS entries, we need to use the service names in DNS.  And thus, we
end up with several PTR entries for 10.0.2.2. :/

Not quite sure how to best adjust this and still get a sensible and
scalable solution.

> I am not an expert regarding that stuff and I don't know if there
> are other ways to achieve the desired. However, it looks as with the
> current setup we need service principals for all host aliases.

That isn't too bad, is it?  It can be added automatically at install
time, right?

Happy hacking,
-- 
Petter Reinholdtsen


-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20110105181024.ga24...@login1.uio.no



Re: NFS4 and Kerberos: A-records for same IP inflate the need for service principals

2011-01-05 Thread Andreas B. Mundt
On Wed, Jan 05, 2011 at 07:10:24PM +0100, Petter Reinholdtsen wrote:
[...]
> > I am not an expert regarding that stuff and I don't know if there
> > are other ways to achieve the desired. However, it looks as with the
> > current setup we need service principals for all host aliases.
> 
> That isn't too bad, is it?  It can be added automatically at install
> time, right?
> 

Yes, the creation of the principals is done during installation. The
script kerberos-kdc-init would contain something like:

for name in tjener.intern kerberos.intern ldap.intern domain.intern 
postoffice.intern syslog.intern; do 
## create machine principals and add them to the keytab:
kadmin.local -q "addprinc -randkey host/$name"
kadmin.local -q "ktadd host/$name"
## create service principals and add them to the keytab:
kadmin.local -q "addprinc -randkey nfs/$name"
kadmin.local -q "ktadd nfs/$name"
kadmin.local -q "addprinc -randkey cifs/$name"
kadmin.local -q "ktadd cifs/$name"
kadmin.local -q "addprinc -randkey ldap/$name"
kadmin.local -q "ktadd -k /etc/krb5.keytab.ldap ldap/$name"
kadmin.local -q "addprinc -randkey imap/$name"
kadmin.local -q "ktadd -k /etc/krb5.keytab.imap imap/$name"
kadmin.local -q "addprinc -randkey smtp/$name"
kadmin.local -q "ktadd -k /etc/krb5.keytab.smtp smtp/$name"
done
chown dovecot:dovecot /etc/krb5.keytab.imap
chown openldap:openldap /etc/krb5.keytab.ldap
chown Debian-exim:Debian-exim /etc/krb5.keytab.smtp

However, I don't know if working with that mess of principals is a
good idea in the end. From a first look it seems like making an
already complicated and hard-to-debug-thing even more confusing, which
also applies to moving individual services to other machines.

Best regards,

 Andi


-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20110105195000.ga12...@flashgordon



Re: NFS4 and Kerberos: A-records for same IP inflate the need for service principals

2011-01-05 Thread Mike Gabriel

Dear Andreas, dear Petter,

On Mi 05 Jan 2011 19:10:24 CET Petter Reinholdtsen wrote:


[Andreas B. Mundt]

I tried to find the reason for these corresponding A-records,


There are two aspects coming together to cause this effect.  We use
strict mode in powerdns, allowing shared A/PTR entries in LDAP, and
the fact that SRV and MX records need to point to A records.  As our
design allow for scaling by moving individual services out by changing
DNS entries, we need to use the service names in DNS.  And thus, we
end up with several PTR entries for 10.0.2.2. :/

Not quite sure how to best adjust this and still get a sensible and
scalable solution.


I am not an expert regarding that stuff and I don't know if there
are other ways to achieve the desired. However, it looks as with the
current setup we need service principals for all host aliases.


That isn't too bad, is it?  It can be added automatically at install
time, right?


Kerberos demands a correct ReverseDNS setup. It can handle multiple  
A-Records for the same IP. Important is that the host principal's IP  
correctly reverse-resolve to the hostname used in the Kerberos host  
principal.


For a correctly working NFS4+Kerberos setup you need (it's quite a  
while ago that I set up my NFS4, so some things might be inaccurate):


  o host principals for all clients and servers of the form:
host/@

so for tjener, this is
host/tjener.int...@intern

and for clients this is
host/dhcp001.int...@intern
...

  o each host has a keytab entry stored locally that corresponds to the host
principal. This keytab hash must be distributed during client installation
(or retrieved via kadmin-server)

  o each service needs a service principal of the form
/@

so these might be:
ldap/tjener.int...@intern
nfs/tjener.int...@intern
etc.

  o these service principals' credentials have to be distributed to all
clients' keytab files individually
on dhcp001: kadmin -q ktadd host/dhcp001.int...@intern
kadmin -q ktadd nfs/dhcp001.int...@intern
on dhcp002: kadmin -q ktadd host/dhcp002.int...@intern
kadmin -q ktadd nfs/dhcp002.int...@intern
...

  o and then you need principals for each user (and pam_krb5.so to get the
tickets for each user on login)

  o some time during squeeze development an extra line in  
/etc/krb5.conf became

necessary:

[libdefaults]
default_realm = INTERN
allow_weak_crypto = true # needed for NFSv4

  o using NFS4+Krb5 starts really making sense when using at least  
sec=krb5i as

mount option. Then you really need a user's Krb5 ticket to be able to
connect to an NFS4 share

  o for this to work you need Kerberos authentication (PAM, see above)
  o and also a running idmapd (nfs-common package)

  o Kerberos5 (MIT) has an LDAP backend, not sure about Heimdal.  
Kerberos is or

was problematic about U.S. export restrictions/regulations. So in this
respect the choice should be Heimdal, but I am not sure if Heimdal has an
LDAP backend...

  o authentication to LDAP via Kerberos can be handled by saslauthd (which is
very neat).

For NFSv4 on Skolelinux I do recommend

  o sec=krb5p for teachers and
  o sec=krb5i for students.

I also recommend one autofs rule for each user and to store them in  
LDAP (nisMapName obejctclass). I am not sure, if Skolelinux's  
automount mechanism already uses this setup. Then at least you can  
store the NFS mount options in LDAP and differentiate between user  
groups concerning nfs4-integrity mounts (sec=krb5i) and nfs4-privacy  
mounts (sec=krb5p).


Mounting all NFS shares via krb5p (or NFS-encrypting one big mount  
point like /skole/home/tjener0) I do absolutely not recommend. NFSv4  
encryption is quite CPU-sucking...


I am really interested in NFSv4+Kerberos5 integration in Skolelinux.  
So if I can be of any help, let me know.


Regards,
Mike

--

DAS-NETZWERKTEAM
mike gabriel, dorfstr. 27, 24245 barmissen
fon: +49 (4302) 281418, fax: +49 (4302) 281419

GnuPG Key ID 0x1943CA5B
mail: m.gabr...@das-netzwerkteam.de, http://das-netzwerkteam.de

freeBusy:
https://mail.das-netzwerkteam.de/freebusy/m.gabriel%40das-netzwerkteam.de.xfb


pgpuFWy7vqmOG.pgp
Description: Digitale PGP-Unterschrift


Your Bugzilla buglist needs attention.

2011-01-05 Thread drift
[This e-mail has been automatically generated.]

You have one or more bugs assigned to you in the Bugzilla bug tracking system 
(http://bugs.skolelinux.org/) that require
attention.

All of these bugs are in the NEW or REOPENED state, and have not been
touched in 7 days or more.
You need to take a look at them, and decide on an initial action.

Generally, this means one of three things:

(1) You decide this bug is really quick to deal with (like, it's INVALID),
and so you get rid of it immediately.
(2) You decide the bug doesn't belong to you, and you reassign it to
someone else. (Hint: if you don't know who to reassign it to, make
sure that the Component field seems reasonable, and then use the
"Reassign bug to default assignee of selected component" option.)
(3) You decide the bug belongs to you, but you can't solve it this moment.
Just use the "Accept bug" command.

To get a list of all NEW/REOPENED bugs, you can use this URL (bookmark
it if you like!):
http://bugs.skolelinux.org/buglist.cgi?bug_status=NEW&bug_status=REOPENED&assigned_to=debian-...@lists.debian.org

Or, you can use the general query page, at 
http://bugs.skolelinux.org/query.cgi

Appended below are the individual URLs to get to all of your NEW bugs
that haven't been touched for a week or more.

You will get this message once a day until you've dealt with these bugs!

 installer ignores mirror/http/proxy preseeding
-> http://bugs.skolelinux.org/show_bug.cgi?id=1458
 ignores mirror/http/hostname preseed
-> http://bugs.skolelinux.org/show_bug.cgi?id=1459


-- 
To UNSUBSCRIBE, email to debian-edu-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/e1paik7-00029r...@maintainer.skolelinux.no