Hi Danishka,

I will try to answer to your questions:

 

Does the controller-1, storage and worker nodes install via PXE server running on the controller-0?

A: Yes. After you install controller-0, it has a PXE server ready to install other nodes. Your initial PXE server is used only for controller-0.

 

When we should set the hostname for controller-1, storage and worker nodes? at the boot prompt?

As per the installation guide, [1] we just need to set the hostname to the new node and set personality at controller-0.

A: from controller-0 only, you set the personality and hostname (if you only need a specific name): system host-update 5 personality=worker hostname=worker-0

 

If controller-0 running a PXE, do we need to upload the boot image?

A: No. controller-0 has everything ready.

 

In this case, if we go with the maximum limit of compute/worker nodes, how we can make the deployment faster?

Do we need to add one node per time?

A: It should work to install in parallel the remaining nodes.

     In our case and I’m talking only about the scripts we are using to automatically install the configurations and check sanity (https://opendev.org/starlingx/test/src/branch/master/automated-robot-suite )

     #1. all servers are off.

     #2. via ipmi we switch on controller-0 server

     #3. it boots from our PXE the bootimage.iso

     #4. after controller-0 is installed (last command: system host-unlock controller-0), via ipmi we switch on all servers

            - on controller-0, with command: system host-list you should see all the powered on servers having None at hostname and personality fields.

     #5. the servers will wait at boot screen for the install script to send proper commands

          The scripts are using a for loop so we are installing sequentially the nodes. We didn’t tried yet to install in parallel because we need to change the current install scripts.

            

 

Regards,

Nicolae Jascanu, Ph.D.

Software Engineer

IOTG

Galati, Romania

 

 

 

From: open infra <openinfradn@gmail.com>
Sent: 2021
41 20:49
To: starlingx-discuss@lists.starlingx.io
Subject: [Starlingx-discuss] Deploying controller-1 and N number of worker nodes

 

Hi,

 

Does the controller-1, storage and worker nodes install via PXE server running on the controller-0?

When we should set the hostname for controller-1, storage and worker nodes? at the boot prompt?

As per the installation guide, [1] we just need to set the hostname to the new node and set personality at controller-0.

If controller-0 running a PXE, do we need to upload the boot image?

 

In this case, if we go with the maximum limit of compute/worker nodes, how we can make the deployment faster?

 

Do we need to add one node per time?

 

 

 

[1] https://docs.starlingx.io/deploy_install_guides/r4_release/bare_metal/dedicated_storage_install_kubernetes.html#install-software-on-controller-1-storage-nodes-and-worker-nodes

 

Regards,

Danishka

---------------------------------------------------------------------------------------------------
Confidentiality Notice: The information contained in this e-mail and any accompanying attachment(s)
is intended only for the use of the intended recipient and may be confidential and/or privileged of
Neusoft Corporation, its subsidiaries and/or its affiliates. If any reader of this communication
is not the intended recipient,unauthorized use,forwarding, printing, storing, disclosure or copying
is strictly prohibited, and may be unlawful.If you have received this communication in error,please
immediately notify the sender by return e-mail, and delete the original message and all copies from
your system. Thank you.
---------------------------------------------------------------------------------------------------