Closed robert-sanfeliu closed 1 month ago
Tested the proof of concept scenario: ProActive jobs: 245,246,247 Failing in nebulous-cd while installing docker on the edge device Success in activeeon environment using scripts: https://github.com/ow2-proactive/scheduling-abstraction-layer/tree/master/docker/scripts
Further validations are to be reported.
tested the scenario with an AWS AMD instance (m7a.large) acting as an edge device registered using the following parameters on the /sal/edge/register
call:
"systemArch": "AMD",
"jarURL": "http://158.39.77.68:8880/rest/node.jar"
Job Ids: 258, 259, 260
Creation of the cluster is successful, including execution of the SAL scripts. Resulting cluster is operational and I can deploy a simple APP
hi Robert, great, thank you for this. Indeed the Intel64 architecture is referred to as ARM. I am improving the error message to notify regarding the supported architectures and will improve the documentation regarding this. Please note that will also need to change this node.jar to be more adapted for Edge. Also, scripts failed on the nebulous probably as the capacity of the AE dummy node is low.
@robert-sanfeliu more complete documentation regarding edge is provided here: https://github.com/ow2-proactive/scheduling-abstraction-layer/blob/master/endpoints/4-edge-endpoints.md
I created a smaller node.jar to be tested for ARM devices: https://www.activeeon.com/public_content/nebulous/node_14.1.0-SNAPSHOT_amd.jar
so can you please test the following that it works fine:
"systemArch": "AMD", "jarURL": "https://www.activeeon.com/public_content/nebulous/node_14.1.0-SNAPSHOT_amd.jar"
Thank you!
tested the scenario with an AWS AMD instance (m7a.large) acting as an edge device registered using the following parameters on the /sal/edge/register call:
"systemArch": "AMD",
"jarURL": "https://www.activeeon.com/public_content/nebulous/node_14.1.0-SNAPSHOT_amd.jar"
Cluster creation was successful.
On nebulous-cd interacting with SAL via REST requests, I'm unable to create a cluster with an edge device with architecture X64. Here are the steps to reproduce:
Register an AWS account
Manually spawn a VM on AWS of type t2.large (uname -a command on the instance shows the architecture:
Linux ip-172-31-19-81 6.5.0-1018-aws #18~22.04.1-Ubuntu SMP Fri Apr 5 17:44:33 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
)grant password based ssh access, set the password and test that it works.
Register the edge node with the following request to SAL (POST to /sal/edge/register):
Get a node candidate for cluster master among the list of available node candidates in the AWS account registered. Select a m4.xlarge instance.
Define a cluster with the following request (POST to /sal/cluster):
Define cluster request returns 200 code and "true" as body.
Call deploy cluster for the defined cluster (POST to /sal/cluster/{{cluster_name}}) , get the following response: