Re[2]: [OpenAFS] OpenAFS client

Ron Croonenberg
Mon, 19 Jul 2004 08:46:37 CT

>Your descriptions of problems are simply too vague to provide much

I am sorry, but I am a bit confused.
It seems that the settings on the server side are the same as they are on an old
cell. and it works fine with the linux client.
The windows client seems to work well with the old cell, but apparently not with
the new cell.

>Switching between cells is therefore unnecessary.  Just make sure your
>CellServDB entries are available OR the AFSDB DNS records are published.

Right,  the CellServDB files should be ok.

>How do you have the Windows client configured?  The last section of
>output from %WINDIR%\TEMP\afsd_init.log would do to describe this?

well there are a few "weird things" in that log.
When it starts everything seems ok it looks like), can find the cellname etc :
7/19/2004 8:23:31 AM: Using >4X4-COWBOY-AFS< as SMB server name
7/19/2004 8:23:31 AM: First Network address a3781272 SubnetMask ffffff00
7/19/2004 8:23:31 AM: rx_SetMaxMTU 1260 successful
7/19/2004 8:23:31 AM: rx_Init code 0
7/19/2004 8:23:31 AM: rx_NewService addr 3a7bd8
7/19/2004 8:23:31 AM: rx_NewService addr 3a7f98
7/19/2004 8:23:31 AM: rx_StartServer
7/19/2004 8:23:31 AM: RPC server listening
7/7/19/2004 8:23:31 AM: cm_InitDCache code 0
19/2004 8:23:31 AM: cm_GetRootCellName code 0, cm_freelanceEnabled= 1, rcn=
7/19/2004 8:23:31 AM: cm_GetSCache code 0 scache c52cf0
7/19/2004 8:23:31 AM: cm_InitDaemon
7/19/2004 8:23:31 AM: smb_localNamep is >4X4-COWBOY-AFS<
7/19/2004 8:23:31 AM: Netbios NCBRESET lana 6 succeeded
7/19/2004 8:23:31 AM: Netbios NCBRESET lana 3 succeeded

then further down the log I see :
7/19/2004 8:23:34 AM: Netbios NCBADDNAME lana=6 code=0 retcode=0 complete=0
7/19/2004 8:23:34 AM: Netbios NCBADDNAME added new name >4X4-COWBOY-AFS  <
7/19/2004 8:23:34 AM: Netbios NCBADDNAME succeeded on lana 6

7/19/2004 8:23:36 AM: Netbios NCBADDNAME lana=3 code=0 retcode=0 complete=0
7/19/2004 8:23:36 AM: Netbios NCBADDNAME added new name >4X4-COWBOY-AFS  <
7/19/2004 8:23:36 AM: Netbios NCBADDNAME succeeded on lana 3

7/19/2004 8:23:36 AM: smb_Init
7/19/2004 8:23:44 AM: New session(ncb_lsn,ncb_lana_num) (1,6) starting from host
4X4-COWBOY, 0 ongoing ops
7/19/2004 8:23:44 AM: increasing numNCBs [ 6 ] numSessions [ 2 ]
7/19/2004 8:23:56 AM: in cm_GetCell_gen cm_SearchCellFile(
returns code= -3 fullname=

And then it seems to have lost the first 3 characters of the name of the cell.
Also, this machine has only one ethernetcard,  but don't the entries for
"lana=3" and "lana=6" indicate there is more then one ?

>Is the following a good summary of your problem?
>    You installed OpenAFS for Windows 1.3.65 with all of the defaults
>    except that you downloaded a custom CellServDB file containing your
>    cell info/ /and specified your.old.cellname as your default cell.
>    Under this configuration everything works fine.  You are able to
>    access the cell, obtain tokens, etc.

correct, I don't have any trouble with the old cell.

>    Now you just installed a new cell.  You modified the default cell
>    name and added new cell server info to CellServDB.  No other changes
>    to the OpenAFS for Windows configuration.   You are no longer able
>    to obtain tokens.

correct, I did all that with the utilities that came with the windows client.

>If so, the questions which come mind are:
>    * which version of OpenAFS for Windows?

the latest one availble.

>    * is KFW installed?  if so, you are using Kerberos 5 for
>      authentication.  Does your new cell have a Kerberos 5 server and
>      did you update your Kerberos 5 configuration or DNS SRV records to
>      allow KFW to obtain tickets.

yes I believe so,  besides.  it works with no problems (that I can see) when
using the linux client instead of the windows client.

>    * if KFW is not installed, you are using Kerberos 4 over UDP via the
>      built-in Kerberos 4 kauth library.  This assumes the KDC is
>      running on the same machines as the AFS servers listed in the
>      CellServDB file.  (ie, kaserver).  Is this in fact running?
>    * Are you able to contact the cell unauthenticated?

I don't know how to do that in windows, but I can contact the cell from
Unix/Linux with no problem

>    * What does "DIR \\afs\cellname" report?

"The network path was not found.

>    * What does "fs listcells" report?  "fs checkservers"?

fs listcells: Cell on hosts

fs checkservers: All servers are running

>    * What about "vos examine root.afs <cellname>"?
C:\>vos examine root.afs
root.afs                          536870927 RW          4 K  On-line /vicepa
    RWrite  536870927 ROnly  536870928 Backup          0
    MaxQuota       5000 K
    Creation    Wed Jul 14 10:02:17 2004
    Last Update Wed Jul 14 11:02:11 2004
    0 accesses in the past day (i.e., vnode references)

    RWrite: 536870927     ROnly: 536870928
    number of sites -> 2
       server partition /vicepa RW Site
       server partition /vicepa RO Site

>Certainly provide more details will help.
>Jeffrey Altman

I can provide you with any details you need.

thank you very much,

I really appreciate it.


     Thomas Edison gets an idea, and his brother Timmy says,
     "Hey, what's that thing over your head?
 Ron Croonenberg               | Phone: 1 765 658 4761
 Technology Coordinator        | Fax:   1 765 658 4732
 Department of ComputerScience | e-mail :
 DePauw University             |
 Julian Science & Math Center  |
 602 South College Ave.        |
 Greencastle, IN  46135        |