Closed oelesinsc24 closed 5 years ago
AFAIK, I have to update the stack with ConfigFramework to deploy a new model. See below:
params=result["Stacks"][0]["Parameters"]
for n,i in enumerate(params):
if(i["ParameterKey"]=="ConfigFramework"):
i["ParameterValue"]="MXNET"
try:
cf.update_stack(
StackName=StackName,
UsePreviousTemplate=True,
Parameters=params,
Capabilities=[
'CAPABILITY_NAMED_IAM',
]
)
waiter = cf.get_waiter('stack_update_complete')
print("Waiting for stack update")
waiter.wait(
StackName=StackName,
WaiterConfig={
'Delay':10,
'MaxAttempts':600
}
)
except ClientError as e:
if(e.response["Error"]["Message"]=="No updates are to be performed."):
pass
else:
raise e
print("stack ready!")
This means that every model training job updates the stack which would hinder a new job once the stack update is in progress. I know a workaround could be to have different stacks for each ConfigFramework
as this helps to isolate problems.
My question is would it be possible to make use of the stack for multiple ConfigFrameworks
at the same time.
Thanks in advance
This question is answered already by deploying multiple stacks.
Thanks @JohnCalhoun
could you explain more about what you would like to see?