[oooq] undercloud-install.sh: No such file or directory

[John Fulton]

quickstart installation fails due to missing undercloud-install.sh on undercloud.

It looks like the "Create undercloud install script" never ran but given the steps I follow to reproduce this (which worked for a while), I don't see why this should be the case.

https://github.com/openstack/tripleo-quickstart-extras/blob/master/roles/undercloud-deploy/tasks/create-scripts.yml#L55-L59

Steps to reproduce:

git clone https://github.com/openstack/tripleo-quickstart.git
cd tripleo-quickstart
export VIRTHOST=$(hostname)
bash quickstart.sh --install-deps
bash quickstart.sh --teardown all --release master-tripleo-ci -e @myconfigfile.yml $VIRTHOST

...

TASK [undercloud-deploy : Install the undercloud] *****************************************************************************************************************************************************************
task path: /home/ooo/.quickstart/usr/local/share/ansible/roles/undercloud-deploy/tasks/install-undercloud.yml:20
Thursday 07 December 2017 00:32:26 -0500 (0:00:06.286) 0:18:37.940 *****
fatal: [undercloud]: FAILED! => {"changed": true, "cmd": "set -o pipefail && /home/stack/undercloud-install.sh 2>&1 | awk '{ print strftime(\"%Y-%m-%d %H:%M:%S |\"), $0; fflush(); }' > /home/stack/undercloud_install.log", "delta": "0:00:00.004757", "end": "2017-12-07 05:32:30.434938", "failed": true, "msg": "non-zero return code", "rc": 127, "start": "2017-12-07 05:32:30.430181", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}

PLAY RECAP ********************************************************************************************************************************************************************************************************
hci-dev.cloud.local : ok=152 changed=72 unreachable=0 failed=0
localhost : ok=17 changed=7 unreachable=0 failed=0
undercloud : ok=3 changed=3 unreachable=0 failed=1

...

[ooo@hci-dev oooq]$ ssh -F ~/.quickstart/ssh.config.ansible undercloud
Warning: Permanently added 'hci-dev.cloud.local,192.168.1.253' (ECDSA) to the list of known hosts.
Warning: Permanently added 'undercloud' (ECDSA) to the list of known hosts.
Last login: Thu Dec 7 05:41:32 2017 from gateway
[stack@undercloud ~]$ cat undercloud_install.log
2017-12-07 05:32:30 | /bin/sh: /home/stack/undercloud-install.sh: No such file or directory
[stack@undercloud ~]$ ls
instackenv.json ironic-python-agent.initramfs ironic-python-agent.kernel overcloud-full.initrd overcloud-full.qcow2 overcloud-full.vmlinuz undercloud_install.log
[stack@undercloud ~]$

[John Fulton (jfulton-org)]

For my reproduction steps, I didn't include that I copied in my myconfigfile.yml into the oooq directory. That files contents are:

[ooo@hci-dev oooq]$ cat myconfigfile.yml
undercloud_memory: 32768 # 12288
undercloud_vcpu: 8 # 4

control_memory: 8192 # 8192
control_vcpu: 2 # 2

compute_memory: 8192 # 8192
compute_vcpu: 2 # 2

ceph_memory: 8192 # 8192
ceph_vcpu: 2 # 2

overcloud_nodes:
  - name: control_0
    flavor: control
    virtualbmc_port: 6230

  - name: control_1
    flavor: control
    virtualbmc_port: 6231

  - name: control_2
    flavor: control
    virtualbmc_port: 6232

  - name: compute_0
    flavor: compute
    virtualbmc_port: 6233

  - name: ceph_0
    flavor: ceph
    virtualbmc_port: 6234

  - name: ceph_1
    flavor: ceph
    virtualbmc_port: 6235

  - name: ceph_2
    flavor: ceph
    virtualbmc_port: 6236
[ooo@hci-dev oooq]$

[Numan Siddique (numansiddique)]

I am able to reproduce the same issue.
Just before the task to deploy undercloud, I addeda task to sleep for a minute and logged into the undercloud VM and created the undercloud-install.sh myself. undercloud install was succcesful. But later failed because some overcloud-prep-*.sh file was missing. Looks like the roles/undercloud-deploy/create-script.yaml playbook is not called at all

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/526334

[Sanjay Upadhyay (saneax)]

Just FYI,

if we use quickstart.sh like below, we hit this bug -
bash quickstart.sh $VIRTHOST

if we use quickstart.sh -t all $VIRTHOST, we are avoiding the scoped tags and do not hit this bug.
the patch adds undercloud-scripts tag to the list, so as by default it is able to run the create-script.yaml.

[John Fulton (jfulton-org)]

Thanks Sanjay, I used that workaround (-t all) and it got me past that point.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/526334
Committed: https://git.openstack.org/cgit/openstack/tripleo-quickstart/commit/?id=1cc9a8eb84e11a7523fc4a9496c1b3bd756f5b25
Submitter: Zuul
Branch: master

commit 1cc9a8eb84e11a7523fc4a9496c1b3bd756f5b25
Author: Sanjay Upadhyay <email address hidden>
Date: Thu Dec 7 16:01:41 2017 +0530

    Fix default deploy where undercloud install scripts are not installed.

    It seems this tag (undercloud-scripts) is not there in default option
    and hence the undercloud-install.sh is not created on the undercloud.

    Refer https://goo.gl/bbbw8j

    Fixes LP Bug #1736859

    Change-Id: I222ca7c3b6743fa3255bd9d2e99285cec18d43d3