[Planetlab-devel] sfi problem

Jeannie Albrecht jeannie at cs.williams.edu
Wed Jul 14 16:08:35 EDT 2010


Thanks Tony, that version seems to work for me.

Jeannie

On Wed, Jul 14, 2010 at 3:41 PM, Tony Mack <tmack at cs.princeton.edu> wrote:
> Hi Jeannie,
>
> It looks like you are using SFA trunk, which uses a different credential and certificate format and is not yet supported by the PlanetLab's public SFA server. Please remove your old certs/creds and try using sfa-0.9-14 instead:
>
> http://svn.planet-lab.org/browser/sfa/tags/sfa-0.9-14
>
> Tony
>
> ----- Original Message -----
> From: "Josh Karlin" <jkarlin at bbn.com>
> To: devel at lists.planet-lab.org
> Sent: Wednesday, July 14, 2010 8:54:20 AM GMT -05:00 US/Canada Eastern
> Subject: Re: [Planetlab-devel] sfi problem
>
>  Ah, okay.  I guess the question is whether or not the server you're
> talking to is updated to the new SFA.  Tony or Scott would have to
> answer that.
>
> Josh
>
>
> On 7/14/10 8:51 AM, Jeannie Albrecht wrote:
>> Hi Josh,
>> I actually did a clean svn checkout yesterday, so the only thing
>> that's "old" is my pkey.  I removed my old certificates and
>> credentials from my ~/.sfi directory.  I'm not running the server or
>> anything...I'm just trying to use the client.  Do I still have to
>> follow those instructions?
>>
>> Jeannie
>>
>> On Wed, Jul 14, 2010 at 8:41 AM, Josh Karlin<jkarlin at bbn.com>  wrote:
>>>   Hi Jeanine,
>>>
>>> Is this an upgrade of SFA over an older version?  If so, the certificates
>>> and creentials have changed.  Please see the upgrade instructions at:
>>> http://groups.geni.net/geni/wiki/UpgradingSFAToTrunk on how to upgrade your
>>> certs/creds.
>>>
>>> Josh
>>>
>>> On 7/14/10 8:03 AM, Jeannie Albrecht wrote:
>>>> Interesting.  I moved my cert and cred files, but I'm still seeing the
>>>> same error.
>>>>
>>>> jeannie at sysnet:~/sfa$ rm /home/jeannie/.sfi/jeannie.cert
>>>> /home/jeannie/.sfi/jeannie.cred
>>>> jeannie at sysnet:~/sfa$ sfi.py list plc.williams
>>>> Traceback (most recent call last):
>>>>    File "/usr/bin/sfi.py", line 1025, in<module>
>>>>      Sfi().main()
>>>>    File "/usr/bin/sfi.py", line 1016, in main
>>>>      self.dispatch(command, cmd_opts, cmd_args)
>>>>    File "/usr/bin/sfi.py", line 552, in dispatch
>>>>      getattr(self, command)(cmd_opts, cmd_args)
>>>>    File "/usr/bin/sfi.py", line 561, in list
>>>>      user_cred = self.get_user_cred().save_to_string(save_parents=True)
>>>>    File "/usr/bin/sfi.py", line 399, in get_user_cred
>>>>      cred.save_to_file(file, save_parents=True)
>>>>    File "/home/jeannie/sfa/sfa/trust/credential.py", line 429, in
>>>> save_to_file
>>>>      self.encode()
>>>>    File "/home/jeannie/sfa/sfa/trust/credential.py", line 378, in encode
>>>>      append_sub(doc, cred, "owner_urn", self.gidCaller.get_urn())
>>>>    File "/home/jeannie/sfa/sfa/trust/credential.py", line 108, in
>>>> append_sub
>>>>      ele.appendChild(doc.createTextNode(text))
>>>>    File "/usr/lib/python2.5/site-packages/_xmlplus/dom/minidom.py",
>>>> line 1608, in createTextNode
>>>>      raise TypeError, "node contents must be a string"
>>>> TypeError: node contents must be a string
>>>>
>>>> Jeannie
>>>>
>>>> On Tue, Jul 13, 2010 at 10:12 PM, Scott Baker<smbaker at gmail.com>    wrote:
>>>>> Ah yes, I remember getting this one too. In my case I think it was due
>>>>> to an old-format credential in my ~/.sfi/ directory. I tracked it to a
>>>>> bug that was occurring when save_to_string() is called on a Credential
>>>>> that was translated from the old format to the new format. Moving the
>>>>> old credential out of the way solved that particular error, but led to
>>>>> another problem with a missing credential signature. Tony mentioned a
>>>>> fix, but I haven't had time to try it out yet. Regardless, I'd be
>>>>> interested to find out if you ran into the same signature issue that I
>>>>> did.
>>>>>
>>>>> Scott
>>>>>
>>>>> On Tue, Jul 13, 2010 at 6:26 PM, Jeannie Albrecht
>>>>> <jeannie at cs.williams.edu>    wrote:
>>>>>> Thanks, Scott.  Unfortunately I'm still getting errors.  Now I see this:
>>>>>>
>>>>>> $ sfi.py list plc.williams
>>>>>> Traceback (most recent call last):
>>>>>>   File "./sfi.py", line 1025, in<module>
>>>>>>     Sfi().main()
>>>>>>   File "./sfi.py", line 1016, in main
>>>>>>     self.dispatch(command, cmd_opts, cmd_args)
>>>>>>   File "./sfi.py", line 552, in dispatch
>>>>>>     getattr(self, command)(cmd_opts, cmd_args)
>>>>>>   File "./sfi.py", line 561, in list
>>>>>>     user_cred = self.get_user_cred().save_to_string(save_parents=True)
>>>>>>   File "/home/jeannie/sfa/sfa/trust/credential.py", line 439, in
>>>>>> save_to_string
>>>>>>     self.encode()
>>>>>>   File "/home/jeannie/sfa/sfa/trust/credential.py", line 378, in encode
>>>>>>     append_sub(doc, cred, "owner_urn", self.gidCaller.get_urn())
>>>>>>   File "/home/jeannie/sfa/sfa/trust/credential.py", line 108, in
>>>>>> append_sub
>>>>>>     ele.appendChild(doc.createTextNode(text))
>>>>>>   File "/usr/lib/python2.5/site-packages/_xmlplus/dom/minidom.py",
>>>>>> line 1608, in createTextNode
>>>>>>     raise TypeError, "node contents must be a string"
>>>>>> TypeError: node contents must be a string
>>>>>>
>>>>>>
>>>>>>
>>>>>> FWIW, my sfi_config looks like this now:
>>>>>>
>>>>>> SFI_AUTH='plc.williams'
>>>>>> SFI_USER='plc.williams.jeannie'
>>>>>> SFI_REGISTRY='http://www.planet-lab.org:12345/'
>>>>>> SFI_SM='http://www.planet-lab.org:12347/'
>>>>>> SFI_GENI_AM='http://www.planet-lab.org:12348/'
>>>>>>
>>>>>> Jeannie
>>>>>>
>>>>>> On Tue, Jul 13, 2010 at 9:13 PM, Scott Baker<smbaker at gmail.com>    wrote:
>>>>>>> I also experienced it here as well. Putting
>>>>>>> SFI_GENI_AM="http://www.planet-lab.org:12348/" in my ~/.sfi/sfi_config
>>>>>>> solved it, although it will also change the default behavior a bit (I
>>>>>>> think it'll start sending slice manager requests to the AM instead of
>>>>>>> the usual SM).
>>>>>>>
>>>>>>> Scott
>>>>>>>
>>>>>>> On Tue, Jul 13, 2010 at 5:49 PM, Jeannie Albrecht
>>>>>>> <jeannie at cs.williams.edu>    wrote:
>>>>>>>> I'm trying to run the newest sfa client and having some strange
>>>>>>>> problems.  Has anyone else encountered this error?
>>>>>>>>
>>>>>>>> $ sfi.py list plc.williams
>>>>>>>> Traceback (most recent call last):
>>>>>>>>   File "/usr/bin/sfi.py", line 1025, in<module>
>>>>>>>>     Sfi().main()
>>>>>>>>   File "/usr/bin/sfi.py", line 1013, in main
>>>>>>>>     self.set_servers()
>>>>>>>>   File "/usr/bin/sfi.py", line 322, in set_servers
>>>>>>>>     self.geni_am = xmlrpcprotocol.get_server(geni_am_url, key_file,
>>>>>>>> cert_file, self.options.debug)
>>>>>>>> UnboundLocalError: local variable 'geni_am_url' referenced before
>>>>>>>> assignment
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Jeannie
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Devel mailing list
>>>>>>>> Devel at lists.planet-lab.org
>>>>>>>> https://lists.planet-lab.org/mailman/listinfo/devel
>>>>>>>>
>>>> _______________________________________________
>>>> Devel mailing list
>>>> Devel at lists.planet-lab.org
>>>> https://lists.planet-lab.org/mailman/listinfo/devel
>>> _______________________________________________
>>> Devel mailing list
>>> Devel at lists.planet-lab.org
>>> https://lists.planet-lab.org/mailman/listinfo/devel
>>>
>> _______________________________________________
>> Devel mailing list
>> Devel at lists.planet-lab.org
>> https://lists.planet-lab.org/mailman/listinfo/devel
>
> _______________________________________________
> Devel mailing list
> Devel at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/devel
>
> _______________________________________________
> Devel mailing list
> Devel at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/devel
>



More information about the Devel mailing list