1

I am trying to install ICP-CE edition on a vagrant provisioned VM. It takes about 1 hour to run and finally fails with the below error.

Let me know if you need any more details or additional error logs

For Version : 3.1.0, Here is the error I am getting. Any help would be appreciated. 

fatal: [localhost]: FAILED! => changed=true
  attempts: 5
  cmd: |-
    filename="/addon/icp-mongodb-3.0.0.tgz"
     if [[ -d "/addon/icp-mongodb-3.0.0.tgz" ]]; then
     filename=$(ls /addon/icp-mongodb-3.0.0.tgz/*.tgz | tail -1)
     fi
     ret=0
     if helm status --tls mongodb &>/dev/null && helm status --tls mongodb | grep -q 'STATUS: FAILED'; then
     ....
               msg: non-zero return code
  rc: 1
  start: '2018-12-12 18:25:55.032453'
  stderr: 'Error: could not find a ready tiller pod'
  stderr_lines: <omitted>
  stdout: |-
    ============================================================
    Tiller log can be found under cluster/logs/tiller-deploy-546cd68bcb-9dv4l
    ============================================================
  stdout_lines: <omitted>

PLAY RECAP *********************************************************************
192.168.17.105             : ok=143  changed=42   unreachable=0    failed=0
localhost                  : ok=112  changed=41   unreachable=0    failed=1

Playbook run took 0 days, 1 hours, 5 minutes, 37 seconds

When I run Version 2.1.0.1, I get the below error - 

TASK [Validating kubelet service] **************************************************************************************************************************************
fatal: [192.168.17.105]: FAILED! => {"changed": false, "failed": true, "msg": "The environment was not clean, please first uninstall the ICP and then reinstall."}

PLAY RECAP *************************************************************************************************************************************************************
192.168.17.105             : ok=26   changed=9    unreachable=0    failed=1

Playbook run took 0 days, 0 hours, 2 minutes, 50 seconds
user6272892
  • 11
  • 2
  • Some basic thoughts: Was the VM "clean" when you started your 3.1.0 installation? Or did you still have 2.1.0.1 installed? Since it seems like the helm tiller might have failed, can you check that the VM has enough resources to run ICP? It could also be helpful to check on the tiller container with `docker ps | grep tiller` – Justin Kulikauskas Dec 13 '18 at 16:00
  • I initially installed a 3.1.* and when it didn't work, I installed a 2.1.*, then got that error. However, I formatted the VM and installed a 3.1.0 all over again after increasing my disk space to 240GB. It failed 4 times and installed successfully on the 5th attempt. I didn't do anything other than hitting the install command again and again on every failure.. Any clue on what caused the failures would be appreciated as I am on a learning journey :) – user6272892 Dec 17 '18 at 17:03
  • The installer uses Helm to deploy other components that ICP needs. If the Helm component goes down, you get the "Error: could not find a ready tiller pod" error you posted. It's not clear what caused Helm to fail, but I would usually suspect the the VM was running out of resources (CPU, RAM), or had some other kind of problem. You might need to look through the logs it mentions in the message to see more details on what went wrong. – Justin Kulikauskas Dec 26 '18 at 16:57

1 Answers1

1

The resource requirements for an 2.1.0.3 ICP install increased. I have installed ICP on a single VM using vagrant, so I know that does work on my 16GB Mac Pro. Given the increased resource requirements for later versions, I am no longer able to run on my machine. You may be running into a hardware resource issue.

Check out the requirements:

Cody Gray - on strike
  • 239,200
  • 50
  • 490
  • 574
Dave Krier
  • 55
  • 6