LFS264: lab1, step 5, APEX setup command fails

eddgest

Hi,
the prerequisite to execute the labs for this course fails: I can login to my google-VM, but when I execute
opnfv-deploy -v --debug -n network_settings.yaml -d os-nosdn-nofeature-noha.yaml
it fails with the following stack-trace:
Traceback (most recent call last):
File "/bin/opnfv-deploy", line 10, in
sys.exit(main())
File "/usr/lib/python3.4/site-packages/apex/deploy.py", line 266, in main
vcpus=args.virt_cpus
File "/usr/lib/python3.4/site-packages/apex/virtual/utils.py", line 68, in generate_inventory
'ipmi_ip': get_virt_ip(),
File "/usr/lib/python3.4/site-packages/apex/virtual/utils.py", line 40, in get_virt_ip
tree = ET.fromstring(virsh_net_xml)
File "/usr/lib64/python3.4/xml/etree/ElementTree.py", line 1335, in XML
parser.feed(text)
xml.etree.ElementTree.ParseError: syntax error: line 1, column 0

It seems there is a parsing error for some file.

Any hints ?
I have attached the apex_deploy.log as well as ther terminal output of my session.

Thanks
Gereon

akapadia

Taking a look

vmuthukrishnan

# This issue is related to openstack controller and compute node memory allocation
# The default ones are not enough, here is the error logs from attached files
#
# 2018-12-10 18:10:31,496 WARNING: RAM per controller is too low.  nosdn requires at least 10GB per controller.
# 2018-12-10 18:10:31,496 INFO: Increasing RAM per controller to 10GB 
#
# Clean OPNFV
opnfv-clean

# Make sure to increase your GCP instance CPU and memory
# Here is the command to increase CPU = 8 (vCPU) & memory = 16 GB   
nohup opnfv-deploy -v --debug -n network_settings.yaml -d os-nosdn-nofeature-noha.yaml --virtual-default-ram 16 --virtual-cpus 8 &


# Kindly note, before starting the above deployment command one need to update the under cloud YUM repository
# It is a know issue in OPNFV Euphrates release (https://jira.opnfv.org/browse/APEX-565)
# Execute the below script, otherwise follow the JIRA ticket workaround to update under cloud YUM repository

sudo -i
cd /opnfv 
nohup ./update_undercloud_yumrepo.sh &

# Wait for the above script to get completed and then start OPNFV deployment using the above command

Let us know, if you face any further issues.

eddgest

Hi,
thanks for looking into this, but it is still happening.
Find the terminal log attached.
regards
Gereon

etodtod

Hello,

I have the same issue with error:

2019-01-11 10:34:03,103 ERROR: Ansible playbook failed. See Ansible logs for details.
Traceback (most recent call last):
File "/bin/opnfv-deploy", line 10, in
sys.exit(main())
File "/usr/lib/python3.4/site-packages/apex/deploy.py", line 339, in main
APEX_TEMP_DIR)
File "/usr/lib/python3.4/site-packages/apex/undercloud/undercloud.py", line 126, in configure
user='stack')
File "/usr/lib/python3.4/site-packages/apex/common/utils.py", line 131, in run_ansible
raise Exception(e)
Exception: Ansible playbook failed. See Ansible logs for details.

I try using the suggested workaround from vmuthukrishnan, but I got the same error.

Attached:

• konsole-output-2019.01.11 => console commands I executed
• nohup-2019.01.11 => printouts from opnfv-deploy -v --debug -n network_settings.yaml -d os-nosdn-nofeature-noha.yaml --virtual-default-ram 16 --virtual-cpus 8
• apex_deploy-2019.01.11

Please help us to fix this issue and be able to proceed with the Labs.

Best regards,
Todor

SriramRupanagunta

Hi,

It appears that you do not have enough vCPUs on the GCP instance to start OPNFV deployment.

gereon@LF-base-machine:~$ gcloud compute instances create opnfv-instance-00 --zone=us-central1-f \

--custom-cpu=8 --custom-memory=52GB --min-cpu-platform="Intel Haswell" \
--image=aarna-opnfv-euphrates-51-pub-05012018-00 --image-project=onap-177920 \
--boot-disk-type "pd-standard" --boot-disk-size=300GB

opnfv-deploy -v --debug -n network_settings.yaml -d os-nosdn-nofeature-noha.yaml --virtual-default-ram 16 --virtual-cpus 8

Your GCP instance has 8 vCPU's, and the OPNFV deployment also uses 8 vCPUs, which will not be sufficient.

Please create GCP instance with 32 vCPUs (this is what we recommend).

Hope that helps.

etodtod

Hello,

Thanks a lot for your replay! I have tried to create a GCP instance with 32 vCPUs, but I got the following error:

gcloud compute instances create opnfv-instance-00 --zone=us-central1-f --custom-cpu=32 --custom-memory=52GB --min-cpu-platform="Intel Haswell" --image=aarna-opnfv-euphrates-51-pub-05012018-00 --image-project=onap-177920 --boot-disk-type "pd-standard" --boot-disk-size=300GB
ERROR: (gcloud.compute.instances.create) Could not fetch resource:
- Quota 'CPUS' exceeded. Limit: 8.0 in region us-central1.

The problem is that as a free GCP user I have limit of 8 vCPUs for all available zones - see attached screen-shoot.

Any suggestions how we can execute the Labs using the free GCP Quotas?

Best regards,
Todor

fcioanca

Unfortunately, the labs cannot be done on a free GCP tier. We actually advise in the course thst some costs are associated with them, due to the complexity of the labs and the environment needed.