swarm-learning icon indicating copy to clipboard operation
swarm-learning copied to clipboard

How to add SN nodes (peers)

Open h-ahmad opened this issue 2 years ago • 4 comments

What about the swop profiles? What about the user, user/app-data folders? What about the increase in PEERS from 2 to 3 in swci? What about the certificates and sharing among peers? Any other change required?

Originally posted by @h-ahmad in https://github.com/HewlettPackard/swarm-learning/discussions/111#discussioncomment-3258135

h-ahmad avatar Jul 27 '22 13:07 h-ahmad

I think you should read documents and try what you think first.

If there is a problem, you can ask here according to the issue template.

IMOKURI avatar Jul 27 '22 22:07 IMOKURI

I think you should read documents and try what you think first.

If there is a problem, you can ask here according to the issue template.

I have gone through the documentation but could not find the procedure to increae the number of SN nodes. I also searched in issues/closed issues and discussion. However, I have tried to add SN node by following changes: Before: After

  1. There were 2 profiles for 2 nodes in swop folder: I have created a third profile with additional index of 2 in each profile.
  2. There were user1 and user2 in first node and user3 & user4 in second node: I have make them up to 3 folders (user1, user2, user3), (user4, user5 user6), (user7, user8, user9) for each 3 nodes.
  3. There was number of PEERS as 2 in swci init file: I have modified to 3.

Remaining settings are same as given in the example. But the problem is that when starting process, two nodes (including sentinel and randomly second; 2 or 3) works well. It does not simply enrol the task for node 2/3. I will attached the screenshot here once I produce that error. Thanks for your response and time.

h-ahmad avatar Jul 28 '22 07:07 h-ahmad

I think you should read documents and try what you think first. If there is a problem, you can ask here according to the issue template.

I have gone through the documentation but could not find the procedure to increae the number of SN nodes. I also searched in issues/closed issues and discussion. However, I have tried to add SN node by following changes: Before: After

  1. There were 2 profiles for 2 nodes in swop folder: I have created a third profile with additional index of 2 in each profile.
  2. There were user1 and user2 in first node and user3 & user4 in second node: I have make them up to 3 folders (user1, user2, user3), (user4, user5 user6), (user7, user8, user9) for each 3 nodes.
  3. There was number of PEERS as 2 in swci init file: I have modified to 3.

Remaining settings are same as given in the example. But the problem is that when starting process, two nodes (including sentinel and randomly second; 2 or 3) works well. It does not simply enrol the task for node 2/3. I will attached the screenshot here once I produce that error. Thanks for your response and time.

Hi, Could you please try with 1.1.0, let us know if issue still exists?

RadhakrishnaJ avatar Aug 07 '22 14:08 RadhakrishnaJ

I think you should read documents and try what you think first. If there is a problem, you can ask here according to the issue template.

I have gone through the documentation but could not find the procedure to increae the number of SN nodes. I also searched in issues/closed issues and discussion. However, I have tried to add SN node by following changes: Before: After

  1. There were 2 profiles for 2 nodes in swop folder: I have created a third profile with additional index of 2 in each profile.
  2. There were user1 and user2 in first node and user3 & user4 in second node: I have make them up to 3 folders (user1, user2, user3), (user4, user5 user6), (user7, user8, user9) for each 3 nodes.
  3. There was number of PEERS as 2 in swci init file: I have modified to 3.

Remaining settings are same as given in the example. But the problem is that when starting process, two nodes (including sentinel and randomly second; 2 or 3) works well. It does not simply enrol the task for node 2/3. I will attached the screenshot here once I produce that error. Thanks for your response and time.

Hi, Could you please try with 1.1.0, let us know if issue still exists?

The version has been updated to 1.1.0, but I have made many changes suggested by the issues/solutions here. For example, container_id issue in swop. Anyway, I will try to implement and let you know if issue persist. Thanks.

h-ahmad avatar Aug 08 '22 07:08 h-ahmad