RSpec fails intermittently at PG Utah

[Sarah Edwards]

Hi Amey,

So interestingly, I tried reserving the tutorial RSpec and 2 out of three times it failed.  But it did eventually work for me.  

So I'd like to suggest that you try again. If it still doesn't work for you after a couple of tries, let us know and we'll investigate further.

Also I'm cc'ing geni-...@googlegroups.com to see if anyone at InstaGENI knows what is happening here.  (You should sign up for the mailing list, making sure to say that you want to receive email, in order to get the responses: https://groups.google.com/forum/#!forum/geni-users)

Cheers,

Sarah Edwards, GPO

On Feb 3, 2014, at 4:01 PM, Amey Jahagirdar <ame...@gmail.com> wrote:

Hello,

I was trying to perform tutorial given here. - http://groups.geni.net/geni/wiki/UDTExampleExperiment

However, create silver request is not getting completed. I tried the same using Flash as well.

Could you please tell me what I am missing?

amey@debian:~$ omni.py createsliver -a pg-utah ajahagirudt http://groups.geni.net/geni/attachment/wiki/UDTExampleExperiment/udt.rspec?format=raw
15:51:13 INFO     omni: Loading agg_nick_cache file '/home/amey/.gcf/agg_nick_cache'
15:51:13 INFO     omni: Loading config file /home/amey/.gcf/omni_config
15:51:13 INFO     omni: Using control framework portal
15:51:13 INFO     omni: Substituting AM nickname pg-utah with URL https://www.emulab.net:12369/protogeni/xmlrpc/am/2.0, URN urn:publicid:IDN+emulab.net+authority+cm
15:51:13 INFO     omni: Slice urn:publicid:IDN+ch.geni.net:netlogistics+slice+ajahagirudt expires on 2014-02-10 20:50:58 UTC
15:51:14 INFO     omni: Creating sliver(s) from rspec file http://groups.geni.net/geni/attachment/wiki/UDTExampleExperiment/udt.rspec?format=raw for slice urn:publicid:IDN+ch.geni.net:netlogistics+slice+ajahagirudt
15:51:56 WARNING  omni: Failed CreateSliver for slice ajahagirudt at https://www.emulab.net:12369/protogeni/xmlrpc/am/2.0.  Error from Aggregate: code 2. protogeni AM code: 2: *** Node mapping precheck failed!*** Node mapping precheck failed!*** ERROR: mapper: Unretriable error. Giving up.
seed = 1391520197
Physical Graph: 205
Calculating shortest paths on switch fabric.
Virtual Graph: 5
Generating physical equivalence classes:59
Type precheck:
Type precheck passed.
Node mapping precheck:
  *** No possible mapping for PC1
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
  *** No possible mapping for PC2
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
*** Node mapping precheck failed!
ASSIGN FAILED:
Type precheck passed.
  *** No possible mapping for PC1
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
  *** No possible mapping for PC2
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
*** Node mapping precheck failed!
 (PG log url - look here for details on any failures: https://www.emulab.net/spewlogfile.php3?logfile=73bb026bf97657fde5512f436a871fdf).
15:51:56 INFO     omni:  ------------------------------------------------------------
15:51:56 INFO     omni:  Completed createsliver:
Args: createsliver ajahagirudt http://groups.geni.net/geni/attachment/wiki/UDTExampleExperiment/udt.rspec?format=raw

  Result Summary: Failed CreateSliver for slice ajahagirudt at https://www.emulab.net:12369/protogeni/xmlrpc/am/2.0.  Error from Aggregate: code 2. protogeni AM code: 2: *** Node mapping precheck failed!*** Node mapping precheck failed!*** ERROR: mapper: Unretriable error. Giving up.
seed = 1391520197
Physical Graph: 205
Calculating shortest paths on switch fabric.
Virtual Graph: 5
Generating physical equivalence classes:59
Type precheck:
Type precheck passed.
Node mapping precheck:
  *** No possible mapping for PC1
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
  *** No possible mapping for PC2
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
*** Node mapping precheck failed!
ASSIGN FAILED:
Type precheck passed.
  *** No possible mapping for PC1
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
  *** No possible mapping for PC2
       OS 'emulab-ops/UBUNTU11-64-STD' (OS-2799) does not run on this hardware type!
*** Node mapping precheck failed!
 (PG log url - look here for details on any failures: https://www.emulab.net/spewlogfile.php3?logfile=73bb026bf97657fde5512f436a871fdf).
15:51:56 INFO     omni:  ============================================================

*******************************************************************************

Sarah Edwards

GENI Project Office

BBN Technologies

Cambridge, MA

phone:    (617) 873-2329

email:    sedw...@bbn.com

[Leigh Stoller]

> So interestingly, I tried reserving the tutorial RSpec and 2 out of three
> times it failed. But it did eventually work for me.
>
> So I'd like to suggest that you try again. If it still doesn't work for
> you after a couple of tries, let us know and we'll investigate further.

Hi. The error message is unfortunately, not very clear. The UBUNTU11 image
runs on a subset of the total nodes in Utah. When all of those nodes are
unavailable (being used by other people), and the only nodes that are
available, are nodes the image cannot run on, the resource mapper prints
that message. In the mapper's view of things, you are trying to use an
image that cannot run on any nodes.

I will have to think about a way to make this more clear ...

Lbs

[Sarah Edwards]

Thanks, Leigh.

Would changing the OS version to a more modern image help?

(I tried a different OS version and it still failed so I had concluded the image wasn't the problem.)

Thanks,
Sarah

> --
> GENI Users is a community supported mailing list, so please help by responding to questions you know the answer to.
>
> If this is your first time posting a question to this list, please review http://groups.geni.net/geni/wiki/GENIExperimenter/CommunityMailingList
> ---
> You received this message because you are subscribed to the Google Groups "GENI Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to geni-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.

[Leigh Stoller]

> Would changing the OS version to a more modern image help?
>
> (I tried a different OS version and it still failed so I had concluded the image wasn't the problem.)

None of the 64 bit images would have worked; all of the 64 bit
capable machines were allocated.

But in general, it is a hard problem in Utah. We have many
many different node types that run run different images cause
of type constraints.

Picking a 32 bit image like FEDORA15-STD would have a better chance
of mapping when things are really tight.

Lbs