Reserve 46 slivers from ProtoGeni

14 views
Skip to first unread message

Chang Liu

unread,
Jun 2, 2015, 12:47:35 PM6/2/15
to protoge...@googlegroups.com
Hi, all,

My experiment needs to reserve 46 slivers (Xen VM) from GENI Aggregate, I have tried a number of times using reservation tool (jFed), but none of them succeeded. I am wondering if anybody has successful experience in reserving these many slivers?  Or is it possible to reserve these many resources at a time? Thanks in advance for any help.

Best
Chang

Hussamuddin Nasir

unread,
Jun 2, 2015, 12:55:18 PM6/2/15
to protoge...@googlegroups.com
We had done this a while back. The trick is to find an Instageni Rack that has very few users using it. Could you spread out the 46 VM onto multiple aggregates and use GENI stitching for L2 links
cheers,

Hussam
(Hussamuddin Nasir)

Netlab Operations Team 

-------------------------------------------------------------------
Laboratory for Adv. Networking  Phone  : (859)218-0059
James F Hardymon Building       Fax    : (859)323-3740
301 Rose Street, Rm 237         E-mail : na...@netlab.uky.edu
Lexington, KY 40506-0495        Web    : http://www.netlab.uky.edu

                        University of Kentucky
                        **********************
------------------------------------------------------------------- 
--
You received this message because you are subscribed to the Google Groups "protogeni-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to protogeni-use...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Chang Liu

unread,
Jun 2, 2015, 1:01:22 PM6/2/15
to protoge...@googlegroups.com
Thanks, Hussam.

My experiment should be ok with stitching, just I heard that stitching was even more easier to fail... I will give a try. Thanks.

Best
Chang 

--
You received this message because you are subscribed to a topic in the Google Groups "protogeni-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/protogeni-users/EN4zCacyT4Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to protogeni-use...@googlegroups.com.

Brecht Vermeulen

unread,
Jun 2, 2015, 1:14:46 PM6/2/15
to protoge...@googlegroups.com

Hi Chang,

it depends on the topology you need (links between the VMs) how you best build this.
You can typically reserve about 10 VMs at once at one aggregate (with raw pcs you can do much more at once), so you can use scale up techniques to try to connect them if needed. (see e.g. mail of Sarah: https://groups.google.com/d/msg/geni-users/KEh88EwZH4Q/47pngtbW-14J )

You can have a look at http://groups.geni.net/geni/wiki/ExpGraphs to see where least XEN VMs are in use.

If it still fails with jFed, let us know (or send a bugreport from the jFed tool)

thanks,
Brecht

Chang Liu schreef op 2/06/2015 om 18:47:

Chang Liu

unread,
Jun 2, 2015, 1:21:29 PM6/2/15
to protoge...@googlegroups.com
Thanks, Brecht. 

I am switching aggregate to a less used one. See if works.

Best
Chang  

--
You received this message because you are subscribed to a topic in the Google Groups "protogeni-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/protogeni-users/EN4zCacyT4Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to protogeni-use...@googlegroups.com.

Chang Liu

unread,
Jun 2, 2015, 4:56:31 PM6/2/15
to protoge...@googlegroups.com
Hi,

I could reserve 46 slivers from a less used aggregate now. Thanks to Hussam and Brecht for the help.

However, after I add the links (there will be 60 links between these 46 nodes in my experiment), it reports error in the first stage when initializing nodes at the aggregate. The error message is:

Create Sliver @ urn:publicid:IDN+geni.kettering.edu+authority+cm: New rspec is not disjoint from sliver.

I am wondering if anybody has idea about what the problem is? Thanks.

Best
Chang

Hussamuddin Nasir

unread,
Jun 2, 2015, 4:58:21 PM6/2/15
to protoge...@googlegroups.com
Did you already reserve slivers at this racks with the same slicename. If so then you may have to delete them first.

cheers,

Hussam
(Hussamuddin Nasir)

Netlab Operations Team 

-------------------------------------------------------------------
Laboratory for Adv. Networking  Phone  : (859)218-0059
James F Hardymon Building       Fax    : (859)323-3740
301 Rose Street, Rm 237         E-mail : na...@netlab.uky.edu
Lexington, KY 40506-0495        Web    : http://www.netlab.uky.edu

                        University of Kentucky
                        **********************
------------------------------------------------------------------- 

Chang Liu

unread,
Jun 2, 2015, 4:59:27 PM6/2/15
to protoge...@googlegroups.com
Yes, I have deleted them. 
Reply all
Reply to author
Forward
0 new messages