Version 3 (modified by 11 years ago) (diff) | ,
---|
Creating and Loading a Disk Image
To prepare his experiment, he loaded a baseline image onto a node on ORBIT's Outdoor.
Try it yourself another time: To load a baseline image onto a node, e.g. node1-6 on ORBIT's Sandbox 4, run
omf load -i baseline.ndz -t node1-3.outdoor.orbit-lab.org
Then he installed all the applications he needs on the node, did any modifications necessary to the coded and saved the disk image.
Try it yourself another time: To save a disk image from a node, e.g. node1-6 on ORBIT's Sandbox 4, run
omf save -n node1-3.outdoor.orbit-lab.org
When the image saving process concluded, it reported that his disk image had been saved and gave the name of the image. Now, when he wants to run his experiment, he will load his disk image onto the nodes he has reserved:
On the client node:
omf load -i gec17_dash_policy_x.ndz -t node1-3.outdoor.orbit-lab.org
Note:
x is either 1, 2, or 3. Here we will implement 3 different policies. These 3 policies are hardwired to the VLC code.
So, we have created 3 different images which need to be loaded when the experimenter needs to test either of them.
On the server node:
omf load -i gec17_dash_apache.ndz.ndz -t node1-5.outdoor.orbit-lab.org
Tip: If any node doesn't seem to be checking in, e.g. you see output like this:
INFO stdlib: Waiting for nodes (Up/Down/Total): 2/1/3 - (still down: node1-5.sb4.orbit-lab.org) [140 sec.]
try opening another SSH session to the outdoor console and resetting the misbehaving node:
omf tell -a reset -t node1-5.outdoor.orbit-lab.org
Clearing Experiment Settings
In addition to loading a fresh disk image, additional steps must be taken to restore the wireless environment to its default state. These commands are run from the console on Sandbox 4 to reset all of the base station and attenuation settings:
wget -qO- "http://wimaxrf:5052/wimaxrf/bs/default" wget -qO- "http://wimaxrf:5052/wimaxrf/bs/restart" wget -qO- "http://wimaxrf:5052/wimaxrf/datapath/config/load?name=default"