| 76 | | Type “step” in the search bar below “Prepare”. It should give options. Select the script ending in “step1-ping_all.rb”. Make any necessary edit to the script so that your node names match(if you used the rspec above the node will already be named properly). Save any changes you made. Then click and drag the paper and pencil icon to the column titled “Execute”.[[BR]] |
| 77 | | [[BR]][[Image(LabwikiTypeStep.png, 50%)]] [[BR]] |
| 78 | | [[BR]][[Image(LabWikiStep1Save.png, 50%)]] [[BR]] |
| 79 | | [[BR]][[Image(LabWikiStep1Drag.png, 50%)]] [[BR]] |
| | 76 | Type “step” in the search bar below “Prepare”. It should give options. Select the script ending in “step1-ping_all.rb”. Click and drag the paper and pencil icon to the column titled “Execute”. [[BR]] |
| | 77 | [[BR]][[Image(LabwikiTypeStep.png, 50%)]] [[BR]] [[BR]] [[Image(LabWikiStep1Drag.png, 50%)]] [[BR]] |
| 81 | | Type your experiment name and slice name in the respective boxes under the Execute column. Write “true” in the graphs box. Then scroll to the bottom of the Execute column and click on “Start Experiment”. Your experiment will begin and you can see progress in the log under the Execute column. When the experiment end a thin green bar will appear in that column.[[BR]] |
| 82 | | [[BR]][[Image(LabWikiStep1Input.png, 30%)]] [[Image(LabWikiStart2.png, 28%)]] [[Image(LabWikiEnd.png, 37%)]] [[BR]][[BR]][[BR]] |
| | 79 | Type your experiment name and slice name in the respective boxes under the Execute column. Make any necessary edits to the boxes labeled source1 through source5 so that they match the names of your nodes(if you used the rspec above the node will already be named properly). Write “true” in the graph box. |
| | 80 | [[BR]][[Image(LabWikiStep1Input.png,40%)]] [[BR]][[BR]][[BR]][[BR]] |
| | 81 | |
| | 82 | Scroll to the bottom of the Execute column and click on “Start Experiment”. Your experiment will begin and you can see progress in the log under the Execute column. When the experiment end a thin green bar will appear in that column.[[BR]] |
| | 83 | [[BR]] [[Image(LabWikiStart2.png, 32%)]] [[Image(LabWikiEnd.png, 43%)]] [[BR]][[BR]][[BR]] |
| 94 | | For GEMINI, the nodes must be named in a specific way. All global nodes must start with the prefix “GN-” and all other nodes must start with the prefix “MP-”. You will want to use an image from a GEMINI ExoGENI rspec for your global node rather than a GIMI image. On your MP nodes include the following section of code to add a postboot script that deals with host name issues: |
| | 95 | For GEMINI, the nodes must be named in a specific way. All global nodes must start with the prefix “GN-” and all other nodes must start with the prefix “MP-”. You will want to use an image from a GEMINI ExoGENI rspec for your global node rather than a GIMI image. Below is an example image for the global node: |
| | 96 | {{{ |
| | 97 | <disk_image name="http://gemini.netlab.uky.edu/Images/vm/exogeni/centos/centos6.3-v1.0.7-GEMINI.xml" version="58b9dd5ff527b570398124443e17ac2e66925bbd"/> |
| | 98 | }}} |
| | 99 | GEMINI needs to execute commands on each node with sudo. Sudo requires that each node has a hostname that is resolvable. Thus we will include a postboot script that sets a unique hostname and adds an entry in the /etc/hosts file so the hostname can be resolved. This allows the sudo commands to run without returning an error code. [[BR]] |
| | 100 | On your MP nodes include the following section of code to add this postboot script: |
| 110 | | In !LabWiki run the four experiment scripts on the slices just as you did with your InstaGENI slices. Instructions are [http://groups.geni.net/geni/wiki/IMInteroperability/SliceWithGIMIAndGEMINI#GIMI here]. Be sure to edit each script so that the names of your nodes match the names in the script. |
| | 116 | In !LabWiki run the four experiment scripts on the slices just as you did with your InstaGENI slices. Instructions are [http://groups.geni.net/geni/wiki/IMInteroperability/SliceWithGIMIAndGEMINI#GIMI here]. Be sure to edit the boxes labelled source1 through source5 so they match the names of your nodes. |
| | 117 | |
| | 118 | Jeanne Ohren: 3. I think it might be good to have a little more detail about what the postboot script is doing. I probably didn't explain that in enough detail. |
| | 119 | Jeanne Ohren: GEMINI needs to execute commands on the node with sudo. Sudo requires that the node has a hostname that is resolvable. The postboot script that you give in the example sets a unique hostname and adds an entry in the /etc/hosts file so the hostname can be resolved. This allows the sudo commands to run without returning an error code. |
| | 120 | Jeanne Ohren: 4. In the section where you mention running the gdesktop scripts and GENI Desktop in dev mode, I think it would be helpful to include examples of how to do that. |
| | 121 | Jeanne Ohren: 5. In the GIMI section where you include the command lines for running the gdesktop scripts, I think it would be helpful to include the last few lines of the output so they know what to look for when the script runs successfully. |
