Cloud Pak for Data

 View Only
  • 1.  AWS Installation Issues

    Posted Thu May 30, 2019 10:48 AM
    Hi, we are trying to install ICP4D on AWS. Having the following issues:

    During the installation, the installer does a LoadBalancer testing. All the tests fail because it takes a minute or so for the LB to become healthy once the service on Port X is up. What can I do to get around this? 

    During the installation, we got the following error:

    TASK [check : Validating /var directory disk space on master nodes] ************

    fatal: [172.31.1.64]: FAILED! => changed=true

      cmd: |-

        disk=$(df -lh -BG --output=avail /var | sed '1d' | grep -oP '\d+')

         [[ $disk -ge 240 ]] || (echo "/var directory available disk space ${disk}GB, it should be greater than or equal to 240 GB" 1>&2 && exit 1)

      delta: '0:00:00.004724'

      end: '2019-05-30 05:04:19.286174'

      msg: non-zero return code

      rc: 1

      start: '2019-05-30 05:04:19.281450'

      stderr: /var directory available disk space 99GB, it should be greater than or equal to 240 GB

      stderr_lines: <omitted>

      stdout: ''

      stdout_lines: <omitted>

    The documentation states that we only need 100GB for the main filesystem.

    During the installation we also encountered the following error:

    TASK [namespace-config : Adding label to system namespaces] ********************

    changed: [172.31.4.12 -> localhost] => (item=kube-system)

    failed: [172.31.4.12 -> localhost] (item=cert-manager) => changed=true

      cmd: |-

        kubectl get namespace cert-manager || kubectl create namespace cert-manager

         kubectl label namespace cert-manager icp=system --overwrite=true

      delta: '0:00:10.261493'

      end: '2019-05-30 05:40:04.282742'

      item: cert-manager

      msg: non-zero return code

      rc: 1

      start: '2019-05-30 05:39:54.021249'

      stderr: |-

        Error from server (NotFound): namespaces "cert-manager" not found

        Unable to connect to the server: net/http: TLS handshake timeout

      stderr_lines: <omitted>

      stdout: namespace/cert-manager created

      stdout_lines: <omitted>

     


    Any idea as to why we got the K8S error? This error basically fails all tasks after this one.

    Thanks in advance

    Guillaume Labelle



    ------------------------------
    Guillaume Labelle
    ------------------------------

    #CloudPakforDataGroup


  • 2.  RE: AWS Installation Issues

    Posted Tue June 04, 2019 02:57 AM
    Hi Guillaume,

    Are you using the "docker_disk" settings in your wdp.conf file? The setting will decide  the docker storage type whether devicemapper of overlay. If you do not plan for devicemapper, then you can comment this setting.

    ------------------------------
    Parthasarthi Komperla
    ------------------------------



  • 3.  RE: AWS Installation Issues

    Posted Tue June 04, 2019 03:50 PM
    Yes I am using that

    ------------------------------
    Guillaume Labelle
    ------------------------------



  • 4.  RE: AWS Installation Issues

    Posted Mon July 29, 2019 12:18 AM
    Edited by System Fri January 20, 2023 04:25 PM

    Common Issues - Amazon Elastic Compute Cloud
    Boot an Amazon EC2 Windows Instance into Directory Services Restore Mode (DSRM) If an instance running Microsoft Active Directory experiences a system failure or other critical issues you can troubleshootthe instance by booting into a special version of Safe Mode called Directory Services Restore Mode (DSRM).(https://docs.aws.amazon.com/AWSEC2/latest/WindowsGuide/common-issues.html)



    ------------------------------
    madhu Dm
    ------------------------------



  • 5.  RE: AWS Installation Issues

    Posted Thu May 18, 2023 09:43 AM

    To address the first issue regarding LoadBalancer testing during installation, you can increase the timeout period for the LoadBalancer health checks. By default, the installer expects the LoadBalancer to become healthy within a short period. However, since your service on Port X takes longer to become healthy, you can adjust the timeout value to accommodate this delay. You can consult the documentation or configuration files of the LoadBalancer service you are using to find the appropriate settings to increase the timeout period.



    ------------------------------
    Eden Wheeler
    ------------------------------