Unable to install openstack

huntingweb

Did anyone install ONAP LAB 2022 after Jan 31st?
I tried but failed.

The error I got:
Error: Failed to download metadata for repo 'appstream': Cannot prepare internal mirrorlist: No URLs in mirrorlist

Quickly google it, found:
CentOS 8 reached its end-of-life on December 31st, 2021, and thus, since January 31st, 2022

What should I do?

ruchirkakkad

That looks like there is no available subscription for Openstack channels. You may run this command to see matching channels for Openstack
Raw

# subscription-manager list --available --matches 'Red Hat OpenStack'

If you don't have required subscription for Openstack then get the subscription, attach it later, and then activate Openstack repos which should work.

huntingweb

Following
https://techglimpse.com/failed-metadata-repo-appstream-centos-8/

I was able to move ahead this early morning.

I did that yesterday, kept getting connection timeout. My guessing is since CentOS 8 reached its end-of-life, all mirrors disappeared and caused the congestion.

I was able to start aio-centos-gcp-deploy.sh, but I ran into another issue:

_ASK [lxc_hosts : Format the machines file] ************************************
TASK [lxc_hosts : Create machines mount point] *********************************
fatal: [aio1]: FAILED! => {"changed": false, "msg": "path /proc/self/fd/4 does not exist", "path": "/proc/self/fd/4
"}
PLAY RECAP *********************************************************************
aio1 : ok=87 changed=4 unreachable=0 failed=1 skipped=9 rescued=0 ignored=0

EXIT NOTICE [Playbook execution failure] **************************************

• grep -q 'Playbook execution success' setup-hosts.out
• echo 'Failed to run the ansible playbook setup-hosts.yml'
  Failed to run the ansible playbook setup-hosts.yml_

Any suggestion?

huntingweb

I was able to fix most of the issues. Still got an error as:
"ImportError: cannot import name 'qos_limit_bandwidth_rules_client'"

According to https://forum.linuxfoundation.org/discussion/comment/32377
We can live with it.

Here are my updates:

update CentOS 8 repository location

Credit goes to https://techglimpse.com/failed-metadata-repo-appstream-centos-8/

#sed -i 's/mirrorlist/#mirrorlist/g' /etc/yum.repos.d/CentOS-*
#sed -i 's|#baseurl=http://mirror.centos.org|baseurl=http://vault.centos.org|g' /etc/yum.repos.d/CentOS-*
#yum update -y

disable selinux

Reference link:
https://docs.openstack.org/openstack-ansible/latest/user/aio/quickstart.html

# systemctl stop firewalld
# systemctl mask firewalld
# vi /etc/sysconfig/selinux
in /etc/sysconfig/selinux, change SELINUX=enforcing is changed to SELINUX=disabled.
# reboot

update CentOS 8 repository location for image used by lxc_hosts

# vi /opt/openstack-ansible/ansible-role-requirements.yml
   comment out following lines, since we need to modify one of file in local
    #- name: lxc_hosts
    #  scm: git
    #  src: https://opendev.org/openstack/openstack-ansible-lxc_hosts
    #  version: fc0eedf9de8a4024b51abbe5771234f89b5bfa32
    #  trackbranch: stable/victoria

#cd /opt/openstack-ansible/playbooks/roles
#git clone https://opendev.org/openstack/openstack-ansible-lxc_hosts lxc_hosts
#cd lxc_hosts
#git checkout stable/victoria
#vi defaults/main.yml 

    change 
      lxc_cache_prep_pre_commands: '## pre command skipped ##'

    to
      lxc_cache_prep_pre_commands: "sed -i 's/mirrorlist/#mirrorlist/g' /etc/yum.repos.d/CentOS-*\n sed -i 's|#baseurl=http://mirror.centos.org|baseurl=http://vault.centos.org|g' /etc/yum.repos.d/CentOS-*"

raghuramg

Hi All,
I will take a look at these errors. If necessary, will create a new image with Ubuntu OS. I will keep you all posted. Thanks.

huntingweb

@raghuramg said:
Hi All,
I will take a look at these errors. If necessary, will create a new image with Ubuntu OS. I will keep you all posted. Thanks.

Do you have any update?

Thanks

raghuramg

Hi @huntingweb ,
Due to CentOS being EndofLife, OpenStack is un-installable on CentOS. There is a warning on OpenStack Page.
We are in process of building a new image with Ubuntu OS. the work is in progress. I will keep you informed here once the image is ready. Thank You for your Patience.

Warning

The announcement of early end-of-life for CentOS 8 and the migration to CentOS Stream leaves OpenStack-Ansible currently un-installable on CentOS Stream. A number of critical dependencies are taken from the EPEL repository and currently these will not install onto a CentOS Stream host. Careful consideration should be given to the life-cycle of any new deployments on CentOS 8.

raghuramg

Hi @huntingweb ,
While we are working on getting a new Ubuntu GCP Image. You could trying installing OpenStack on any VM with Ubuntu 20.04 OS which has a public IP. You could follow the same steps as in the LAB guide.

Copy the onap-custom folder from the GCP to target ubuntu VM where you will install OpenStack and update the script (~/onap-custom/setup-utils/aio-centos-client-install.sh) with below content and try.

===================================================================

!/bin/bash

LOGFILE="aio-openstack-cli-installation.log"

echo "Going to install the Openstack Cli and Heat Stack Cli packages"

apt install python3-pip -y | tee -a $LOGFILE
python3 -m pip install -U pip | tee -a $LOGFILE
pip3 install python-openstackclient | tee -a $LOGFILE
pip3 install python-heatclient | tee -a $LOGFILE

echo "Successfully installed the Client packages"

exit 0

raghuramg

Hi @huntingweb,
Were you able to install OpenStack on any other VM other than GCP ?

huntingweb

I am still waiting your office image.
I had stuck OpenStack issue for 3 days. Really don't want to wast any more time on it.
Currently, I temporarily used an old version of microstack, it is super easy, supports http and didn't see any issue in OOM installation.

sudo snap install microstack --classic --channel=rocky/edge

Do you have the official version ready?

raghuramg

Hi @huntingweb,
The new OpenStack image is not available yet. We're having some issues with installing OpenStack on GCP. We are debugging the issue. Hopefully, the image will be available by next week. I will keep you informed.

huntingweb

@raghuramg said:
Hi @huntingweb,
The new OpenStack image is not available yet. We're having some issues with installing OpenStack on GCP. We are debugging the issue. Hopefully, the image will be available by next week. I will keep you informed.

Hi @raghuramg,
Any update on this?

Thanks

raghuramg

Hi @huntingweb,
We've fixed the OpenStack issue on Ubuntu. The new GCP image with Ubuntu OS will be available by tomorrow. I will comment here once it is available.

huntingweb

Hi @raghuramg

Since there were a few issues found during recent exercise, can you guy re-certify the lab after fixing the Openstack installation issue and AAI certification problem?

raghuramg

Hi @huntingweb,
The new GCP Ubuntu OpenStack Image is Ready.

Here are the steps to install with a new image.
# Launch an instance to deploy OpenStack
gcloud compute instances create openstack-01 --zone us-central1-f --custom-cpu 32 --custom-memory 72GB --min-cpu-platform "Intel Haswell" --image "https://www.googleapis.com/compute/v1/projects/onap-177920/global/images/aarna-aio-openstack-victoria-ubuntu-20-16032022-00"

#If you see this error for utility container. Please ignore.
aio1_utility_container-35e97cfd : ok=57 changed=3 unreachable=0 failed=1 skipped=26 rescued=0 ignored=0

# Replace the following attached scripts in path ~/onap-custom/setup-utils
Extract the attached scripts.zip to get the updated files.
aio-centos-gcp-deploy.sh
aio-centos-client-install.sh

Follow the remaining steps as per the lab document to install OpenStack.

# Find the ip address of your eth0 interface
ip addr show ens4

seanch

@raghuramg
I'm trying to install Openstack using the updated Ubuntu image, however it is currently hanging on:

RUNNING HANDLER [os_keystone : Wait for uWSGI socket to be ready] **************

In the nohup.log I see a bunch of messages saying backend has no server available:

Message from [email protected] at Mar 17 17:56:27 ...
 haproxy[104364]: backend nova_api_os_compute-back has no server available!

Broadcast message from [email protected] (Thu 2022-03-17 17:56:27 UTC):

haproxy[104364]: backend nova_console-back has no server available!


Message from [email protected] at Mar 17 17:56:27 ...
 haproxy[104364]: backend nova_console-back has no server available!

Broadcast message from [email protected] (Thu 2022-03-17 17:56:28 UTC):

haproxy[104364]: backend placement-back has no server available!


Message from [email protected] at Mar 17 17:56:28 ...
 haproxy[104364]: backend placement-back has no server available!

Broadcast message from [email protected] (Thu 2022-03-17 17:56:29 UTC):

haproxy[104364]: backend rabbitmq_mgmt-back has no server available!

Is there something else that needs to be configured?

raghuramg

Hi @seanch,
You can ignore backend has no server available messages.
Could you please share your VM public IP? I would like to take a look at it.

michaober

Hi,

the install fails for me. Please see the attached logfile. I'm using the Ubuntu image and replaced the scripts.

raghuramg

Hi @michaober,
Have you created VM using the below command? If yes, Could you please share your VM Public IP ?

gcloud compute instances create openstack-01 --zone us-central1-f --custom-cpu 32 --custom-memory 72GB --min-cpu-platform "Intel Haswell" --image "https://www.googleapis.com/compute/v1/projects/onap-177920/global/images/aarna-aio-openstack-victoria-ubuntu-20-16032022-00"

michaober

I created it on "us-central1-a", not "f", because "f" had no resources at the time I tried.

I have sent you the IP via private message.

raghuramg

Hi @michaober,
I cannot SSH or Ping the VM IP that you shared. Please verify.

michaober

I've started it again and will leave it running for the next few hours.

raghuramg

Hi @michaober,
I could not connect to the VM still. Please make sure the IP is correct?

michaober

The IP is correct, but I cannot leave the VM running 24/7 due to costs. It was running for ~8 hours after my previous post.

There has to be another way to solve this issue.

raghuramg

Hi @michaober,
I am in India TimeZone. So I have missed your message last night. Is it possible to create a new VM with the below command now and share the IP?

gcloud compute instances create openstack-01 --zone us-central1-f --custom-cpu 32 --custom-memory 72GB --min-cpu-platform "Intel Haswell" --image "https://www.googleapis.com/compute/v1/projects/onap-177920/global/images/aarna-aio-openstack-victoria-ubuntu-20-16032022-00"

raghuramg

Hi @michaober,
I have started the installation and it is in progress. so far no errors. it is going smoothly. I will update once it gets complete. it might take around the next 2 hours.

raghuramg

Hi @michaober,
It has failed again at the below steps. I could not clean up completely to start the install again. is it possible to delete this VM and create a new VM?

TASK [os_keystone : Create database for service] *******************************
failed: [aio1_keystone_container-21a606d1 -> 172.29.237.184] (item=None) => {"censored": "the output has been hidden due to the fact that 'no_log: true' was specified for this result", "changed": false}
fatal: [aio1_keystone_container-21a606d1]: FAILED! => {"censored": "the output has been hidden due to the fact that 'no_log: true' was specified for this result", "changed": false}

michaober

I'm pretty sure that repeating the same steps while get the same results.

Can you maybe create a working VM and document the steps you took?

raghuramg

Hi @michaober,
It is a working Image. I recommend not to power-off ONAP or OpenStack VM's even when not used. That might cause problems. we have seen issues before if power off and power on.

Have you tried on a new VM ? or this is the 1st VM ?

michaober

I recommend not to power-off ONAP or OpenStack VM's even when not used. That might cause problems. we have seen issues before if power off and power on.

How is this supposed to work? Is there a way to use GCP for free? If powering off and on breaks the system, I would not consider it is "working".

Have you tried on a new VM ? or this is the 1st VM ?

It's the first. But if I'm supposed to spend another working day on getting this system up and running, I'd like to have more concise instructions. Currently, I picked everything from different sources - the PDF supplied in the course and different forum posts.

So I'd like to ask you again to provide complete, end-to-end, tested instructions.

raghuramg

Hi @michaober,

Here are the steps to install OpenStack. The image is tested and it will work fine if VM is not restarted / Power-off.

# Launch an instance to deploy OpenStack
gcloud compute instances create openstack-01 --zone us-central1-f --custom-cpu 32 --custom-memory 72GB --min-cpu-platform "Intel Haswell" --image "https://www.googleapis.com/compute/v1/projects/onap-177920/global/images/aarna-aio-openstack-victoria-ubuntu-20-16032022-00"

#If you see this error for utility container. Please ignore.
aio1_utility_container-35e97cfd : ok=57 changed=3 unreachable=0 failed=1 skipped=26 rescued=0 ignored=0

cd onap-custom/
git pull -v
cd setup-utils/
nohup sudo bash -x aio-ubuntu-gcp-deploy.sh 2 &
sudo bash -x ./aio-ubuntu-update-haproxy-http.sh
ip addr show ens4
sudo bash -x ./aio-prepare-overcloudrc.sh aarna