perfSONAR User Q&A and Other Discussion

[Mark Yashar <>]

Hi Mark,

I'm trying to run a throughput test via the perfsonar-toolkit (versoin 4.0.0.2) web interface.

Can you let me know how I would be able to tell if the test completed successfully, and

how within the the web interface I can find the results and graphs of the results? According to the

perfSONAR documentation, "Plots of perfSONAR test results can be viewed by selecting

Details on the Toolkit web interface homepage". I wasn't able to find this and I'm not sure

how I can tell if the test was run successfully and was complete.

Thanks and best regards,

- Mark Yashar

On Sat, Jul 29, 2017 at 7:19 PM, Vera Wang <> wrote:

Hi Mark,

Thank you for your advice and help. We'll try installing Centos and then the perfsonar bundle and see how it works.

Thanks,

Vera

On Thursday, July 27, 2017, Mark Feit <> wrote:

Vera Wang writes:

 

We think it's because upon installing, the version of Centos Perfsonar uses defaults to a custom install without a GUI like Gnome, and we've tried to manually install using the GUI install versus a minimal install but it doesnt allow us to continue with the installation process after selecting an option.

 

The toolkit is designed as an appliance, so what we install is fairly stripped down.  The usual MO as described in the installation guide is to install the software on a machine and access it remotely with a browser running on another.

 

If you want the GUI and browser on the same machine, a full CentOS install and the perfSONAR bundle will get you there the same way.

 

Another weird thing that happens is that from the ssh we are unable to ping, and from what I understand don't have network access. Also our version of Perfsonar custom centos doesn't have the right repo for the yum command upon trying to install a GUI that way. I could find and send you the error messages if that would help pinpoint the problem.

 

If you don’t have network access, YUM isn’t not going to work since the CentOS repos are out on the Internet.  We don’t provide the full image.  So the network is the first thing to resolve.  When you installed, you should have been given the option to configure the network interface. The default behavior is to use DHCP, so if that’s not available where you’re plugged in, it’s going to end up a non-starter.  I’ve never installed Linux on a Mac, but from what I understand, the network hardware is bog standard and shouldn’t be a problem.

 

I’d suggest running through the installation again and making sure the network is properly configured.  If the installer doesn’t ask, the hardware wasn’t recognized, and that’s about the limit of where I can help you out.

 

--Mark

 

--
Class of 2020

--

Mark Yashar

Specialist, UC Berkeley

CITRIS and the Banatao Institute

citris-uc.org