Top level HPCC-Charms repository for build and deployment. All the related projects are just prototypes. They are open for re-design and re-implementation. There two groups of charms: 1) Single HPCC Platform which has all supported plugins and module. The initial implementation includes layer-hpccsystems-platform and layer-hpccsystems-plugins 2) HPCC Cluster The initial implementation includes layer-hpccsystems-cluster-node and layer-hpccsystems-cluster-manager. Since both groups involved multiple charms we hope a group of charm bundles will be provided to easy deployment of various HPCC Charm setups. Each deployable charm also should have test using amulet APIs.

• Ubuntu 18.04 amd64 Xenial
• Python 3.0
• juju-2.5" sudo snap install juju
• lxd : sudo snap install lxd
• charm-tools: sudo snap install charm

• git clone https://github.com/hpcc-charms/HPCC-Charms.git
• git submodule update --init
• create a build directory
• cd to HPCC-Charms and run ./build.sh. The build output will be in ../build directory

#lxd init --auto (probably already done) #lxd init --auto --storage-backend zfs (probably already done) lxc network set lxdbr0 ipv6.address none

#Check network: lxc network get lxdbr0 ipv4.address #Sample output: 10.172.37.1/24

 lxc network get lxdbr0 ipv4.address
 juju bootstrap localhost lxd-test
 juju bootstrap aws aws-test

juju models | grep default
[ $? -ne 0 ] && juju add-model default
juju deploy <build dir>/builds/hpcc-platform hpcc --series bionic
juju status

cd  <build dir>/trusty/hpcc-cluster-dali
mv config.yaml config.yaml.orig
cp config_node.yaml config.yaml

juju models | grep default
[ $? -ne 0 ] && juju add-model default
juju deploy ./hpcc-cluster-dali dali --series bionic
juju deploy ./hpcc-cluster-node node -n 2 --series bionic
juju status
#juju debug-log

When 3 units (1 dali 2 nodes) are started add relation

juju add-relation dali node

When 3 units in "joined" states run followint to stop/create envxml/fetch/start HPCC:

juju config dali update-envxml=$(date +"%T.%6N")

This uses date but any different value in "update-envxml" will work If everything run OK all unit should be in "HPCC.started"

The esp will be on unit: node/0

You can get ip to test EclWatch: http://:8010

juju debug-log

For non install error you can login the unit to change code

juju ssh <unit>
#go to dir to fix code
cd /var/lib/juju/agents/unit-<unit>/charm/

In another local windows:

juju resovled <unit>

This will allow the hook re-run

juju destroy-model default

##Troubleshooting

juju kill-controller ERROR getting controller environ: unable to get bootstrap information from client store or API Solution: juju unregister -y