Skip to content

Latest commit

 

History

History
94 lines (76 loc) · 4.39 KB

QA-PROCESS.md

File metadata and controls

94 lines (76 loc) · 4.39 KB

QA Process

Create stack

Follow the directions to bring up the stack. After the first-time initialization and creation of a virtual environment it should run like this:

NAME=vis-qa
source activate refinery-deploy

cd refinery-platform/deployment/terraform/live/
terraform workspace select default
terraform workspace delete -force $NAME
# ... but ideally it would be clean and not need -force? 
terraform workspace new $NAME
TF_LOG=DEBUG terraform apply

# Needs to be run from the terraform directory to fill in the template:
erb ../../aws-config/config.yaml.erb > ../../aws-config/config.yaml

cd ../..
perl -i -pne '$_="ADMIN_PASSWORD: admin-password" if /ADMIN_PASSWORD:/' aws-config/config.yaml
perl -i -pne '$_="SSH_USERS: mccalluc jkmarx scottx611x hackdna" if /SSH_USERS:/' aws-config/config.yaml
python stack.py create

# Wait for it to complete:
while true; do aws cloudformation describe-stacks --stack-name $NAME --query 'Stacks[*].[StackStatus]' --output text; sleep 2; done

... or keep an eye on the AWS console.

Update DNS: (Because of a virtual host setting or something like that, you can not hit the server directly: It needs to have the right hostname and go through ELB.)

ELB_DNS=dualstack.`aws elb describe-load-balancers --load-balancer-names $NAME --query 'LoadBalancerDescriptions[*].DNSName' --output text`
ZONE_ID=`aws elb describe-load-balancers --load-balancer-names $NAME --query 'LoadBalancerDescriptions[*].CanonicalHostedZoneNameID' --output text`
echo '{"Changes": [{"Action": "UPSERT", "ResourceRecordSet": {"Name": "'$NAME'.cloud.refinery-platform.org", "Type": "A", "AliasTarget": {"HostedZoneId": "'$ZONE_ID'", "DNSName": "'$ELB_DNS'", "EvaluateTargetHealth": false}}}]}' > /tmp/dns.json
aws route53 change-resource-record-sets --hosted-zone-id Z1D2YVM2HNPAQB --change-batch file:///tmp/dns.json

(Note that this uses two different Zone IDs: on the commandline, it the zone for our own apex domain, which is different from the domain for the ELB DNS.)

Set up

At this point, puppet may still be running. Log in, tail the log, and wait for Cloud-init v. x.y.z finished:

IP=`aws ec2 describe-instances  --filters Name=tag:Name,Values=$NAME --query 'Reservations[].Instances[].PublicIpAddress' --output=text`
ssh -i ~/.ssh/mccalluc.pem ubuntu@$IP tail -f /var/log/cloud-init-output.log

After puppet completes, create a user account on the instance:

ssh -i ~/.ssh/mccalluc.pem ubuntu@$IP 
cd /srv/refinery-platform/refinery
workon refinery-platform
USERNAME=vis-qa
./manage.py create_user $USERNAME $USERNAME-password [email protected] first last hms True

Install tools:

# Continue in ssh...
# TODO: List from branch, and load from branch.
GIT_API_URL=https://api.github.com/repos/refinery-platform/visualization-tools/contents/tool-annotations
TOOLS=`python -c 'import requests; print(" ".join([tool["name"].replace(".json","") for tool in requests.get("'$GIT_API_URL'").json()]))'`
./manage.py load_tools --visualizations $TOOLS

Load sample data:

TODO: Do it in the UI until refinery-platform/refinery-platform#2946 is fixed

# Continue in ssh...
wget https://raw.githubusercontent.com/refinery-platform/django_docker_engine/master/demo_path_routing_no_auth/upload/demo-data.csv
./manage.py process_metadata_table --username vis-qa --title vis-qa-data --file_name demo-data.csv --source_column_index 1 --data_file_column 0 --delimiter comma

Do the QA

Visit http://vis-qa.cloud.refinery-platform.org/ and login with vis-qa/vis-qa-password.

For each tool:

  • Load all the appropriate data.
  • Excercise all the functionality.
  • Load some inappropriate data and make sure it fails gracefully.

If there are problems...

  • regressions from the currently deployed state need to be fixed before release
  • but other bugs can just be filed, probably in the repo for the appropriate tool.

Destroy stack

aws cloudformation delete-stack --stack-name vis-qa
while aws cloudformation describe-stacks --stack-name vis-qa > /dev/null; do echo 'still up'; sleep 10; done
for ID in `aws ec2 describe-snapshots --filters Name=tag:aws:cloudformation:stack-name,Values=vis-qa --query 'Snapshots[].[SnapshotId]' --output=text`; do aws ec2 delete-snapshot --snapshot-id $ID; done