Uploaded image for project: 'Help-Desk'
  1. Help-Desk
  2. HELP-8452

[Fiware-lab-help] FIWARE Lab Assistance

    Details

    • Type: extRequest
    • Status: Closed
    • Priority: Major
    • Resolution: Done
    • Fix Version/s: 2021
    • Component/s: FIWARE-TECH-HELP
    • Labels:
      None

      Description

      Dear Sir / Madam,

      I am looking for advices and support for the installation of a new assiociated node for the Haute École Arc Ingénierie organization.

      We did all steps in https://forge.fiware.org/plugins/mediawiki/wiki/fiware/index.php/FIWARE_Lab_Nodes_Handbook for the installation of the new OpenStack environment using OPS-Deploy.
      We did it on VirtualBox for testing purpose, and now we are interested in a real installation.

      Since we have a limited amount of resources for this node, we do not want a production-ready environment with high-availability for 100+ users but a sufficient environment for ~10 working VMs with some Fiware Docker components on them.

      Do you think that the following architecture would be enough for the requirements above:

      1 Master Node

      • 2 CPU
      • 2 GB RAM
      • 60 GB HDD

      1 Slave Node as a Controller

      • 2 CPU
      • 2 GB RAM
      • 60 GB HDD

      1 Slave Node as a Compute

      • 2 CPU
      • 2 GB RAM
      • 60 GB HDD

      1 Slave Node as a Storage

      • 2 CPU
      • 2 GB RAM
      • 500 GB HDD

      Any suggestions are welcomed, even if you drastically increase the required resources.

      I also have another question : is the OPS-Deploy-v4.0 still the preferred way (with default configurations) for the deployment of a future Fiware Lab Node or is there a new, more recent document that I am missing?

      Thank you in advance for your answer. You can also tell me if you want me to ask the question in http://stackoverflow.com/questions with the [fiware] tag.

      Yours sincerely
      Dany Jupille

      __________________________________________________________________________________________

      You can get more information about our cookies and privacy policies on the following links:

      Fiware-lab-help mailing list
      Fiware-lab-help@lists.fiware.org
      https://lists.fiware.org/listinfo/fiware-lab-help

      [Created via e-mail received from: Jupille Dany <dany.jupille@he-arc.ch>]

        Activity

        Hide
        danieles Daniele Santoro added a comment -

        Hi Fernando,

        if I am not wrong, according to the FI-Next roadmap, OPS-Deploy-v4.0 should not be used anymore. Every region should install OpenStack using the preferred way, then after the federation process FI-Deploy new tools should be used just to install the FI-Health components. Probably a warning about the EOL of OPS-Deploy-v4.0 should be added to the README. I can ask to Federico and provide it.

        About the resources needed, in my opinion they should be increased drastically, but you know better than me.

        Best,d

        Show
        danieles Daniele Santoro added a comment - Hi Fernando, if I am not wrong, according to the FI-Next roadmap, OPS-Deploy-v4.0 should not be used anymore. Every region should install OpenStack using the preferred way, then after the federation process FI-Deploy new tools should be used just to install the FI-Health components. Probably a warning about the EOL of OPS-Deploy-v4.0 should be added to the README. I can ask to Federico and provide it. About the resources needed, in my opinion they should be increased drastically, but you know better than me. Best,d
        Hide
        fla Fernando Lopez added a comment -

        Dear Dany,

        Like it was mentioned by Daniele, the installation of a node basically follows the OpenStack manuals in that case you do not need to use a Master node to keep the Ops-Deploy tool. Due to currently you do not plan to federate your node, you do not need anything else especial to install the node, probably the list of Generic Enablers images available in FIWARE to play with them.

        Regarding the configuration, taking into account that you want to have just only 10 VMs the controller and storage node should be enough. Nevertheless the compute node is very poor in resources. You have to keep in mind that the flavour that you probably required will have 2 GB RAM, therefore you will require at least 24 GB RAM if you plan only one computer (this is the minimum that I suggest to you to work, probably slowly). If you change the Master node to a Compute node, in that case you will required 14 GB on each machine. You have to keep in mind that RAM is usually the bigger problems in virtualization environments due to you cannot virtualize it.

        In case of disk capacity, you can select Virtual Machines without disk, but you have to keep in mind that the image of the Virtual Machine should be a file, and usually (depends of the format of this file) it is around 9 GB of data, therefore you need to store around 90GB data if you planned to use 10 VMs with minimum capacity. If you use 2 Compute nodes with the description that you provided, you can do it. Nevertheless, keep in mind that you have logs files, OS files and so on, probably you do not have capacity for 5 VMs only 4 VMs, it is something that you see on the way.

        I think that it is enough for the moment.

        Best regards,

        Fernando

        Show
        fla Fernando Lopez added a comment - Dear Dany, Like it was mentioned by Daniele, the installation of a node basically follows the OpenStack manuals in that case you do not need to use a Master node to keep the Ops-Deploy tool. Due to currently you do not plan to federate your node, you do not need anything else especial to install the node, probably the list of Generic Enablers images available in FIWARE to play with them. Regarding the configuration, taking into account that you want to have just only 10 VMs the controller and storage node should be enough. Nevertheless the compute node is very poor in resources. You have to keep in mind that the flavour that you probably required will have 2 GB RAM, therefore you will require at least 24 GB RAM if you plan only one computer (this is the minimum that I suggest to you to work, probably slowly). If you change the Master node to a Compute node, in that case you will required 14 GB on each machine. You have to keep in mind that RAM is usually the bigger problems in virtualization environments due to you cannot virtualize it. In case of disk capacity, you can select Virtual Machines without disk, but you have to keep in mind that the image of the Virtual Machine should be a file, and usually (depends of the format of this file) it is around 9 GB of data, therefore you need to store around 90GB data if you planned to use 10 VMs with minimum capacity. If you use 2 Compute nodes with the description that you provided, you can do it. Nevertheless, keep in mind that you have logs files, OS files and so on, probably you do not have capacity for 5 VMs only 4 VMs, it is something that you see on the way. I think that it is enough for the moment. Best regards, Fernando

          People

          • Assignee:
            fla Fernando Lopez
            Reporter:
            fw.ext.user FW External User
          • Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved: