Questions & Answers

Bringing up PNDA on a bare metal server with 6VMs

0 votes
asked Apr 26 by kingpinH (320 points)
edited Apr 26 by kingpinH
I have a server with 6VMs (as per the pico requirements) with nothing but salt master installed on one of the nodes.

The machine that I have is a bare-metal machine with 32 cores and 96 GB RAM and ample storage.

Once I configure the other nodes as minions and create this salt stack cluster, will running the pnda-cli take care of all of the provisioning?

1 Answer

0 votes
answered Apr 26 by jeremie (3,130 points)
Thanks for your question.

Could you clarify your server configuration in term of hardware and if this is OpenStack or not? We provide cli which automate everything, including the infrastructure provisioning and the deployment, no need to install and configure a Salt master, this is part of the process. If you want to have mode details, check https://github.com/pndaproject/pnda-guide/tree/master/provisioning

If this is bare metal, with a set of nodes, this is currently on the develop branch and will be part of the next PNDA release 3.5, but do not think this is something you are looking for.

Cheers,

Jérémie
commented Apr 26 by kingpinH (320 points)
My machine is not provisioned using Openstack if that is what you are asking. It is an independent machine.
Would installing devstack on it ease my life?
commented May 2 by trsmith2 (1,740 points)
PNDA can be provisioned on OpenStack, AWS or bare metal. Please choose one of these and make a platform ready before proceeding with the installation process. The resources you'll need to make available are described here - https://github.com/pndaproject/pnda-guide/blob/develop/provisioning/platform_requirements.md

So for example if you choose to deploy on OpenStack you'll need to make available through OpenStack sufficient resources to support at least 18 vcpus 53 GB memory and 195 GB storage. It's not important how you do this - we do not mandate DevStack, for example - but in the case of OpenStack you must make sure you choose a version of OpenStack that is supported (Kilo or above) and install all the services (including Swift and Heat).
commented May 2 by yusong0926 (240 points)
can you provide a recommended local.conf file for devstack?
commented 4 days ago by kingpinH (320 points)
# Allow running on not-supported linux distro/version
FORCE=false
OFFLINE=false
# Reclone repositories everytime devstack runs
RECLONE=false
GIT_BASE=https://git.openstack.org

LOGFILE=$DATA_DIR/devstack.log
SCREEN_LOGDIR=$DATA_DIR/logs

FLOATING_RANGE=192.168.123.0/24
Q_FLOATING_ALLOCATION_POOL="start=192.168.123.133,end=192.168.123.200"
PUBLIC_NETWORK_GATEWAY=192.168.123.248
#HOST_IP=192.168.2.100

Q_USE_SECGROUP=True
ENABLE_TENANT_VLANS=True
TENANT_VLAN_RANGE=1000:1999
PHYSICAL_NETWORK=public
OVS_PHYSICAL_BRIDGE=br-ex


# Credentials
ADMIN_PASSWORD=password
MYSQL_PASSWORD=password
RABBIT_PASSWORD=password
SERVICE_PASSWORD=password
SERVICE_TOKEN=password

# Services
enable_service rabbit

## Enable Neutron
disable_service n-net
disable_service tempest
enable_service q-svc
enable_service q-agt
enable_service q-dhcp
enable_service q-l3
enable_service q-meta
enable_service neutron
#enable_service tempest
enable_service horizon
# Enable heat
enable_service h-eng h-api h-api-cfn h-api-cw
# Enable ceilometer
enable_plugin ceilometer https://git.openstack.org/openstack/ceilometer.git stable/ocata


# Cinder configuration
# ========================================
CINDER_SECURE_DELETE=false
# Two minions by default will create 25GB size volumes, so make your that we have anough space for volumes
VOLUME_BACKING_FILE_SIZE=100000M

# Enabling Swift
# ========================================
enable_service s-proxy s-object s-container s-account
SWIFT_HASH=66a3d6b56c1f479c8b4e70ab5c2000f5
SWIFT_REPLICAS=1
SWIFT_DATA_DIR=/opt/stack/data/swift
...