Changes between Version 37 and Version 38 of Tutorials/g0WmLTE/Tutorial2

Timestamp:

Jul 19, 2012, 3:01:20 PM (12 years ago)

Author:

jeffrabi

Comment:

—

Tutorials/g0WmLTE/Tutorial2

@@ -115 +115 @@
 Optionally, double check that the image loaded correctly.
 
-Now, launch the experiment!
+Now, turn the nodes on and launch the experiment!
 
 {{{
+omf tell on all
 omf-5.3 exec nyupoly_mcsmod.rb
 }}}
@@ -128 +129 @@
 There are a few problems which may arise during the course of attempting to run this experiment. Usually, they are the result of fragmentation of framework/operational code. Below are some errors encountered, and suggested solutions.
 
- * The node handler gives an error stating that the experiment expected an integer/string and that it was given a string/integer. This arises due to different versions of the source ''iperf.rb'' application wrapper. Simply edit the Experiment Description to the alternate data type and try again.
-
-[[Image(datatypeerror.png)]]
-
-
  * The node handler gives an error stating the network profile failed to configure, but then the raised error describes an ongoing connection status -- this is most likely a false alarm caused by running an experiment over a predefined configuration. The node reconfigures, but believes that the configuration status is an error message. 
 
 [[Image(profileerror.png)]]
 
+
+ * The experiment "hiccups" and groups/nodes cannot be accessed or configured -- sometimes this is nothing more than a hiccup in accessing a node. However, occasionally it means you've forgotten to turn the nodes on. The worst-case scenario is that the nodes are experiencing connectivity issues. The best way to confirm it is the latter and not the former is to return to the previous experiment and confirm whether it is possible to ping the nodes. (They need to be manually configured despite having the experiment script run partially.)
 ----
 
@@ -156 +154 @@
 sqlite3 -init myDatabase myDatabase.sq3
 }}}
+
+----
+
+== Final Observations == 
+
+Due to the highly experimental nature of ORBIT and its OMF and OML, it is possible that this experiment may eventually be broken by future design changes. Please bear this in mind.
+
+If this experiment does not work for you, and the above troubleshooting suggestions did not resolve your difficulties, you may need to contact an administrator. Unfortunately, it is difficult to tell the difference between the system being down and simply misconfigured for an end user. Therefore, you should attempt to debug the experiment using the above troubleshooting tips to ensure that the error is not from the configuration.
+
+----