Azure
/
azureml-examples
зеркало из https://github.com/Azure/azureml-examples.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

List resource before validation to avoid not-found error (#3365) 

* list resource before checking

* list resource before checking

---------

Co-authored-by: Ubuntu <zetia@DevBox-zetia.1jltvvkrfgyuhl3llhmbyldkog.bx.internal.cloudapp.net>
This commit is contained in:
Zeliang Tian 2024-09-12 11:23:55 +08:00 коммит произвёл GitHub
Родитель 0833654898
Коммит efe9b8cf33
Не найден ключ, соответствующий данной подписи
Идентификатор ключа GPG: B5690EEEBB952194
1 изменённых файлов: 25 добавлений и 14 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

39
infra/bootstrapping/sdk_helpers.sh
Просмотреть файл

@ -812,7 +812,7 @@ function replace_version(){
function ensure_k8s_compute(){
# Arc cluster configuration
arc_compute=${ARC_CLUSTER_NAME}
echo_info "Creating amlarc cluster: '$arc_compute'"
echo_info "Checking amlarc cluster: '$arc_compute'"
# Remove AKS if unhealthy
compute_exists=$(az aks list --resource-group "${RESOURCE_GROUP_NAME}" --query "[?name == '${arc_compute}']" |tail -n1|tr -d "[:cntrl:]")
@ -826,23 +826,34 @@ function ensure_k8s_compute(){
fi
# Remove Arc if unhealthy
clusterState=$(az connectedk8s show --resource-group "${RESOURCE_GROUP_NAME}" --name "${arc_compute}-arc" --query connectivityStatus -o tsv)
echo_info "Cluster: ${arc_compute}-arc current state: ${clusterState}"
if [[ "${clusterState}" != "Connected" ]]; then
echo_info "Remove unhealthy ARC: '${arc_compute}-arc'"
az connectedk8s delete --resource-group "${RESOURCE_GROUP_NAME}" --name "${arc_compute}-arc" --yes
compute_exists=$(az connectedk8s list --resource-group "${RESOURCE_GROUP_NAME}" --query "[?name == '${arc_compute}-arc']" |tail -n1|tr -d "[:cntrl:]")
if [[ "${compute_exists}" = "[]" ]]; then
echo_info "Arc Compute ${arc_compute}-arc does not exist; will create"
else
clusterState=$(az connectedk8s show --resource-group "${RESOURCE_GROUP_NAME}" --name "${arc_compute}-arc" --query connectivityStatus -o tsv)
echo_info "Cluster: ${arc_compute}-arc current state: ${clusterState}"
if [[ "${clusterState}" != "Connected" ]]; then
echo_info "Remove unhealthy ARC: '${arc_compute}-arc'"
az connectedk8s delete --resource-group "${RESOURCE_GROUP_NAME}" --name "${arc_compute}-arc" --yes
fi
fi
# Remove k8s compute if unhealthy
Status=$(az ml compute show --resource-group "${RESOURCE_GROUP_NAME}" --name "${ARC_COMPUTE_NAME}" --query provisioning_state --output tsv)
if
[[ $Status == "Succeeded" ]]
then
echo_info "Compute is healthy: $Status"
compute_exists=$(az ml compute list --resource-group "${RESOURCE_GROUP_NAME}" --query "[?name == '${ARC_COMPUTE_NAME}']" |tail -n1|tr -d "[:cntrl:]")
if [[ "${compute_exists}" = "[]" ]]; then
echo_info "K8s Compute ${arc_compute}-arc does not exist; will create"
else
echo_info "Compute is unhealthy: $Status"
az ml compute detach --subscription "${SUBSCRIPTION_ID}" --resource-group "${RESOURCE_GROUP_NAME}" --workspace-name "${WORKSPACE_NAME}" --name "${ARC_COMPUTE_NAME}" -y || true
fi
Status=$(az ml compute show --resource-group "${RESOURCE_GROUP_NAME}" --name "${ARC_COMPUTE_NAME}" --query provisioning_state --output tsv)
if
[[ $Status == "Succeeded" ]]
then
echo_info "K8s Compute is healthy: $Status"
else
echo_info "K8s Compute is unhealthy: $Status"
az ml compute detach --subscription "${SUBSCRIPTION_ID}" --resource-group "${RESOURCE_GROUP_NAME}" --workspace-name "${WORKSPACE_NAME}" --name "${ARC_COMPUTE_NAME}" -y || true
fi
fi
LOCATION=eastus2 ensure_aks_compute "${arc_compute}" 1 3 "STANDARD_D3_V2"
install_k8s_extension "${arc_compute}" "connectedClusters" "Microsoft.Kubernetes/connectedClusters"