Creting fat-tree topology using "OpenFlow OVS all XEN" RSpec

[Sadia Bashir]

Hello everyone,

I successfully performed my experiment using OpenFlow OVS all XEN RSpec. Now I want to extend this experiment to a Fat-tree topology and want to create fat-tree topology using the same node and link specification as is used by this RSpec. I tried creating Fat-tree topology by duplicating this RSpec's nodes and using Clemson InstaGENI Aggregate but it fails every time. I tried creating fat-tree topology using MAX InstaGENI aggregate as well but it also fails with error. I also tried creating binary tree topology with 15 nodes which completes with no error.  Please see the files attached for topology and error and guide me figuring out what I am doing wrong and how can it be fixed. I'll be really grateful for your time, help or suggestions.

Looking forward seeing your response.

Kind Regards,
Sadia Bashir

[Nicholas Bastin]

On Thu, Aug 13, 2015 at 2:55 AM, Sadia Bashir
<11msccs...@seecs.edu.pk> wrote:
> which completes with no error. Please see the files attached for topology
> and error and guide me figuring out what I am doing wrong and how can it be
> fixed. I'll be really grateful for your time, help or suggestions.

In the future please send actual text rather than screenshots so we
can copy/paste the URLs instead of typing them by hand
(http://www.instageni.maxgigapop.net/spewlogfile.php3?logfile=265a7ffda4129f89df8e6221cc0b26da
if anyone is still looking...)

How did you create this rspec? The problem you're having at the
moment is that you have a lot of <service> tags that don't contain any
commands, which is apparently not a legal rspec format, so the AM is
rejecting it.

--
Nick

[Sarah Edwards]

Hi Sadia,

Two other comments to add to Nick's message:
* UtahDDC InstaGENI is the site with the most resources so that would be a good location for trying to reserve this topology in the future.
* You can use the rspeclint [1] tool to validate that your rspec conforms to the specification. If you used a graphical tool (e.g. Jacks in the portal or jFed) to write the tutorial and you find that an RSpec generated by the tool does not pass rspeclint, that would definitely be a bug we would want to know about. If that is the case, please send the request RSpec to this list with a note about the tool used to create it and the error found by rspeclint.

Thank you,
Sarah


[1] http://www.protogeni.net/wiki/RSpecDebugging

> --
> 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/d/optout.

*******************************************************************************
Sarah Edwards
GENI Project Office

BBN Technologies
Cambridge, MA
phone: (617) 873-2329
email: sedw...@bbn.com

[Sadia Bashir]

 Thank you so much for your prompt response.

How did you create this rspec?  

    I created it using GENI Portal Jack tool and specifying "OpenFlow OVS all XEN" rspec by duplicating its host and ovs nodes.

[Sadia Bashir]

Two other comments to add to Nick's message:
 * UtahDDC InstaGENI is the site with the most resources so that would be a good location for trying to reserve this topology in the future.
 * You can use the rspeclint [1] tool to validate that your rspec conforms to the specification. If you used a graphical tool (e.g. Jacks in the portal or jFed) to write the tutorial and you find that an RSpec generated by the tool does not pass rspeclint, that would definitely be a bug we would want to know about.  If that is the case, please send the request RSpec to this list with a note about the tool used to create it and the error found by rspeclint.

 

 Hi Sarah,
 Thank you so much for your suggestions. Really appreciated.

Thank you,
Sadia

[Sarah Edwards]

Can you send us the request RSpec?

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/d/optout.

[Nicholas Bastin]

At least one problem is in your node "agg-1", which has an execute
script set, but with no command actually provided.

--
Nick

[Sarah Edwards]

Hi,

This is caused by a bug in Jacks in the Portal. I have reported the bug to the maintainers of that code.

The workaround for now is to remove the empty execute and install fields from the "agg-1" node (either in Jacks or by manually deleting the fields).

Cheers,
Sarah

[Sadia Bashir]

This is caused by a bug in Jacks in the Portal.  I have reported the bug to the maintainers of that code.

The workaround for now is to remove the empty execute and install fields from the "agg-1" node (either in Jacks or by manually deleting the fields).

 

Thank you so much Sarah, Thank you so much Nicholas, for all your support and guideline. I created the topology once again, used UtahDDC InstaGENI aggregate and it was successful. 

Thank you,

Sadia

[Sadia Bashir]

Hello everyone,

I am getting the following run status error after the fat tree topology has been created successfully using UtahDDC InstaGENI aggregate:

Error with two attempts:

1. Nodes:
  [Node: pcvm8-11]
in urn:publicid:IDN+ch.geni.net:N

extGenSDN+slice+sdnHelloHWS failed.

pcvm8-11        : https://www.utahddc.geniracks.net/spewlogfile.php3?logfile=23011f87e3cfb1aabec1c5b6839dbc45

 sfdisk: ERROR: sector 0 does not have an msdos signature
 /dev/xen-vg/pcvm8-11: unrecognized partition table type

2. Nodes:
  [Node: pcvm7-18]
in urn:publicid:IDN+ch.geni.net:NextGenSDN+slice+sdnHelloHWS failed.

pcvm7-18        : https://www.utahddc.geniracks.net/spewlogfile.php3?logfile=8c14a0cd215c8716b0543b871e3a08bf

 sfdisk: ERROR: sector 0 does not have an msdos signature
 /dev/xen-vg/pcvm7-18: unrecognized partition table type

I have attached my RSpec file for this topology as well, In each attempt, only one but different node failed to start with the error specified above.
I'll be really grateful if someone here can guide me to fix it.

Thank you,
Sadia Bashir

[Leigh Stoller]

> 1. Nodes:
> [Node: pcvm8-11]
> in urn:publicid:IDN+ch.geni.net:N
> extGenSDN+slice+sdnHelloHWS failed.
>
> pcvm8-11 : https://www.utahddc.geniracks.net/spewlogfile.php3?logfile=23011f87e3cfb1aabec1c5b6839dbc45

Well, it is kinda strange. This particular VM shows this on the console:

[ 16.593416] ------------[ cut here ]------------
16.593428] kernel BUG at /build/buildd/linux-3.13.0/arch/x86/kernel/paravirt.c:239!
[ 16.593434] invalid opcode: 0000 [#1] SMP

The reason it is strange is that this is the UBUNTU14-OVS2.31 image, which appears
to boot okay on other nodes in your topology?

Leigh