Replicate the experiment with ccpe-2014

103 views
Skip to first unread message

Ricardo Almeida

unread,
Jan 11, 2014, 7:57:14 PM1/11/14
to opensta...@googlegroups.com
Hello, I read the thread [How to replicate the experiment] and I have some questions about the tests. I perceived that the repository cited in the old thread has moved to https://github.com/beloglazov/ccpe-2014-experiments. So, like to confirm that the procedures for the tests remain the same. And if the image file ubuntu2 referenced in the script spe-vms-boot-28-slow.sh there's somewhere?

Sorry, my English isn't good. Finally congratulations, your thesis was excellent.

Anton Beloglazov

unread,
Jan 13, 2014, 10:04:00 PM1/13/14
to opensta...@googlegroups.com
Hi Ricardo,

Thank you for your interest in the project! You are right, the repository has moved to the location you pointed out. The procedures are the same as before. Unfortunately, the image is not available for download, please follow the steps in the following message that describe how to configure a VM image: https://groups.google.com/d/msg/openstack-neat/PKz2vpKPMcA/-eImY47z1rMJ

Best regards,
Anton

Ricardo Almeida

unread,
Jan 23, 2014, 7:02:40 PM1/23/14
to opensta...@googlegroups.com
Hi Anton, I’m sorry for the delay to reply. I was very busy, but now, I have progress with the tests.

Unfortunately I am having problem with script workload-distribuitor.py, after 28 requests of computes I get IndexError (empty list) because of pop. The workload-traces that I am testing is equal (full-utilization), so I declared the dir and files variables inside of poll function. For my initial tests I believe that it will be reasonable (correct?), but do you have any suggestion to solution this problem?

Thanks. Best reguards.

Anton Beloglazov

unread,
Jan 28, 2014, 4:13:15 AM1/28/14
to opensta...@googlegroups.com
Hi Ricardo,

How exactly did you specify the path to the full-utilization directory? BTW, the path in the command variable (lines 40-41) also needs to be adjusted according to your setup. The problem you describe might also happen if you have more than 28 VMs requesting for workloads, that's why when all 28 traces have been used, it fail to serve the 29th request.

Cheers,
Anton


--
You received this message because you are subscribed to the Google Groups "OpenStack Neat" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openstack-nea...@googlegroups.com.
Visit this group at http://groups.google.com/group/openstack-neat.
For more options, visit https://groups.google.com/groups/opt_out.

Reply all
Reply to author
Forward
0 new messages