Newer
Older
variables:
GIT_SUBMODULE_STRATEGY: recursive
NECTAR_ALLOCATION: HPCCICD
stages:
# - integration_test_downstream # working but unwanted here
# - trigger_pipeline_in_B # working but unwanted here
- heat
- ansible_create_cluster_stage
- push_button_spawn_cluster
# - e2e
- tests
yamllint:
allow_failure: true
tags:
- yamllint
script:
- echo "stage yamllint"
- cd CICD
- yamllint -c ./.yamllintheat.yaml ./heat
# delete_stack_manual:
# stage: delete_stack_manual
# tags:
# - heat
# before_script:
# - echo "$GC_KEY" > gc_key.pem
# - chmod 400 gc_key.pem
# - echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
# script:
# - echo "heat stage"
# - source ./$NECTAR_ALLOCATION-openrc.sh
# - openstack stack list
# - bash -x ./CICD/heat/heatcicdwrapper.sh delete_if_exists $STACKNAME
# when: manual
ansiblelint:
Luhan Cheng
committed
allow_failure: true
script:
- echo "stage ansiblelint"
- cd CICD
- python3 ansiblelint/run_lint.py --targets master_playbook.yml
build_cluster_cicd:
stage: heat
allow_failure: false
tags:
- heat
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
script:
- echo "heat stage"
- source ./$NECTAR_ALLOCATION-openrc.sh
- openstack stack list
- bash -x ./heat/heatcicdwrapper.sh create_or_update $STACKNAME
- bash -x ./heat/server_rebuild.sh all
after_script:
- sleep 30 # artifically wait a bit to give the nodes time to boot
# only:
# changes: #https://docs.gitlab.com/ee/ci/yaml/#onlychangesexceptchanges
# - "heat/*HOT*.yaml"
# - schedules
# - ./.gitlab-ci.yml
ansible_create_cluster_stage:
stage: ansible_create_cluster_stage
tags:
- ansible
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
script:
- echo "ansible_create_cluster_stage"
- bash -x ./CICD/ansible_create_cluster_script.sh
- ansible-playbook -i files/inventory.$STACKNAME --key-file ../gc_key.pem master_playbook.yml
- sleep 15
- echo uglyuglyfix
- ansible -i files/inventory.$STACKNAME --key-file ../gc_key.pem -b -a "systemctl restart slurmdbd" ManagementNodes
- sleep 60
- echo do it again
- ansible-playbook -i files/inventory.$STACKNAME --key-file ../gc_key.pem master_playbook.yml
tests:
stage: tests
tags:
- ansible
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
script:
- echo "tests stage"
- source ./$NECTAR_ALLOCATION-openrc.sh
- openstack stack list
- cd CICD
- python3 ../scripts/make_inventory.py static $STACKNAME | tee ./files/inventory.$STACKNAME && chmod 755 ./files/inventory.$STACKNAME
- grep -qv "I could not find any resouces tagged with project_name:" ./files/inventory.$STACKNAME #fail if inventory file is empty
- ansible -m ping -i files/inventory.$STACKNAME --key-file ../gc_key.pem all
- ansible -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "sudo ls" all
- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "sinfo" ManagementNodes
- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "squeue" ManagementNodes
- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet mariadb" SQLNodes
- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet slurmctld" ManagementNodes
- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet slurmdbd" ManagementNodes
- bash -e ./tests/run_tests.sh all "files/inventory.$STACKNAME" "../gc_key.pem"
- bash -e ./tests/run_tests.sh ComputeNodes "files/inventory.$STACKNAME" "../gc_key.pem"
- bash -e ./tests/run_tests.sh LoginNodes "files/inventory.$STACKNAME" "../gc_key.pem"
- bash -e ./tests/run_tests.sh ManagementNodes "files/inventory.$STACKNAME" "../gc_key.pem"
- bash -e ./tests/run_tests.sh NFSNodes "files/inventory.$STACKNAME" "../gc_key.pem"
- bash -e ./tests/run_tests.sh SQLNodes "files/inventory.$STACKNAME" "../gc_key.pem"
extended:
stage: extended
tags:
- heat
- ansible
before_script:
- echo "cleanup stack"
- sleep 30
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
- source ./$NECTAR_ALLOCATION-openrc.sh
- bash -x ./CICD/heat/heatcicdwrapper.sh delete_if_exists $STACKNAME
only:
variables:
- $EXTENDED != null
manual_cluster_spawn:
stage: push_button_spawn_cluster
tags:
- heat
- ansible
before_script:
- echo "press button spawn cluster."
- echo "for this to work you have to provide a variable called manual stackname"
- echo I still need to handle os password
- echo $MANUAL_STACKNAME
- echo "$GC_KEY" > gc_key.pem
- chmod 400 gc_key.pem
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
script:
- source ./$NECTAR_ALLOCATION-openrc.sh
- bash -x ./CICD/heat/heatcicdwrapper.sh create $MANUAL_STACKNAME
- openstack stack list
- export STACKNAME=$MANUAL_STACKNAME
- sleep 25
- bash -x CICD/ansible_create_cluster_script.sh
when: manual
only:
refs:
- "cicd"
clean:
stage: clean
tags:
- heat
before_script:
- echo "cleanup stack"
- sleep 30
- echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
script:
- source ./$NECTAR_ALLOCATION-openrc.sh
- bash -x ./CICD/heat/heatcicdwrapper.sh delete_if_exists $STACKNAME
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
#trigger_pipeline_in_B:
# stage: integration_test_downstream
# tags:
# - ansible
# script:
# - "curl --request POST --form token=${CI_JOB_TOKEN} --form ref=master https://gitlab.erc.monash.edu.au/api/v4/projects/1085/trigger/pipeline" # ID is from pysshauthz
# heat_test:
# stage: heat_test
# allow_failure: false
# tags:
# - heat
# before_script:
# - echo "$GC_KEY" > gc_key.pem
# - chmod 400 gc_key.pem
# - echo "$HPCCICD_openrc" > ./$NECTAR_ALLOCATION-openrc.sh
# - source ./$NECTAR_ALLOCATION-openrc.sh
# - export HEAT_TEST_STACKNAME=_TESTING_HEAT
# - bash -x ./CICD/heat/heatcicdwrapper.sh delete_if_exists $HEAT_TEST_STACKNAME
# - sleep 60
# script:
# - echo "heat_test stage"
# - source ./$NECTAR_ALLOCATION-openrc.sh
# - bash -x ./CICD/heat/heatcicdwrapper.sh create $HEAT_TEST_STACKNAME
# - openstack stack list
# - bash -x ./CICD/heat/heatcicdwrapper.sh update $HEAT_TEST_STACKNAME
# - openstack stack list
# - bash -x ./CICD/heat/heatcicdwrapper.sh delete_if_exists $HEAT_TEST_STACKNAME
# - openstack stack list
# after_script:
# - sleep 20 # artifically wait a bit to make sure it is really dead