Closed chuckwondo closed 1 year ago
I was able to do a successful sandbox deploy, but not yet able to run the smoke test.
Making a note here about this deployment in case it helps anyone as a future reference.
It was a little bit of a bumpy road but here is what I experienced. Commands used 'make pre-deploy-setup' // 2 times 'make all-up-yes' // multiple times 'make up-cumulus' // To enter EarthData Credentials 'make up-cumulus-yes' // multiple times to solve timing based errors
Here is a kind of chronology of what the deployment.
(1) make pre-deploy-setup
Result
#181 terraspace init cumulus]: ERROR: Aws::Errors::MissingCredentialsError: unable to sign request without credentials set
#181 terraspace init cumulus]: It doesnt look like AWS credentials and access has been setup.
Error running: terraspace init cumulus. Fix the error above or check logs for the error.
Resolution
I double checked .env files and credentials - everything looked good so I ran it again.
(2) make pre-deploy-setup
Result
Successful but still had these 'errors' listed
terraspace init rds-cluster: ERROR: Aws::Errors::MissingCredentialsError: unable to sign request without credentials set
terraspace init rds-cluster: ERROR: Aws::Errors::MissingCredentialsError: unable to sign request without credentials set
terraspace init rds-cluster: Terraform has been successfully initialized!
Resolution
I went ahead to the next step of running 'make all-up-yes'
(3) make all-up-yes
Result
Error: error creating Lambda Function (1): InvalidParameterValueException: The provided execution role does not have permissions to call CreateNetworkInterface on EC2
Resolution
Ran 'make all-up-yes' again to solve the deployment racing conditions.
(4) make all-up-yes
Result
EarthData Login Error
On this run, there was never an obvious pause or prompt to enter the credentials.
I think the 'yes' part of the command causes that.
Resolution
Ran 'make up-cumulus' to
(5) make up-cumulus
Result
After the Earthdata login, the deploy timed out waiting for a 'yes'
Resolution
Ran 'make up-cumulus-yes' // Now that we are past the Earth Data login, this should work fine.
(6) make up-cumulus-yes
Result
This part of the deploy took a long time.
For the most part, it ran smoothly and created mosst of the services.
At one point I did have to manually remove a lock with these commands
make bash
terraspace force_unlock cumulus <LOCK_GUID> // The deployment script gives you this command
At another point I got a resource already exists error and had to remove it with this command.
make terraform-doctor-cumulus
The final result of the deploy were these lines
archive_api_redirect_uri = https://t9maohtgsi.execute-api.us-west-2.amazonaws.com:8000/dev/token
archive_api_uri = https://t9maohtgsi.execute-api.us-west-2.amazonaws.com:8000/dev/
cumulus_distribution_api_redirect_uri = login
cumulus_distribution_api_uri =
report_executions_sns_topic_arn = arn:aws:sns:us-west-2:772632177894:cumulus-kris-sbx7894-report-executions-topic
report_granules_sns_topic_arn = arn:aws:sns:us-west-2:772632177894:cumulus-kris-sbx7894-report-granules-topic
report_pdrs_sns_topic_arn = arn:aws:sns:us-west-2:772632177894:cumulus-kris-sbx7894-report-pdrs-topic
stepfunction_event_reporter_queue_url = https://sqs.us-west-2.amazonaws.com/772632177894/cumulus-kris-sbx7894-sfEventSqsToDbRecordsInputQueue
Note:
The reason the line above cumulus_distribution_api_uri =
was blank is because I did not comment out the line TF_VAR_cumulus_distribution_url=
in the .env file for this fresh deployment.
When setting up the .env file for a FIRST deploy to any given environment, the line TF_VAR_cumulus_distribution_url=
Should be commented out like this
#TF_VAR_cumulus_distribution_url=
Re: The Smoke Test - Stretch goal
Here is a link to the description of why the smoke test was not able to be completed. https://github.com/NASA-IMPACT/csdap-cumulus/pull/115#issuecomment-1430046463
Attempted to run the smoke test in a less automated way by running these two commands.
make bash aws s3 cp --recursive "app/stacks/cumulus/resources/granules/PSScene3Band" "s3://csda-cumulus-kris-sbx7894-provider-7894/storage-ss-ingest-prod-ingesteddata-uswest2/planet/" >/dev/null
The granules do make it up to the bucket that way, but when attempting to run the rules, another NoSuchBucket error pops up when trying to load the granules from the provider bucket.
Smoke Tests passed after incorporating the changes in this Pull Request: https://github.com/NASA-IMPACT/csdap-cumulus/pull/116
Standup Cumulus 13.4.0 deployment to
csda-app-sbx-7894
CBA AWS account for Kris to use for development, by following the instructions in this repo'sREADME.md
file.Acceptance criteria:
README.md
instructions, with the exception of running a successful E2E test (i.e., all resources managed by Terraform are successfully created)README.md
file (if this cannot be completed in the same sprint as the item above, this issue should still be closed, and new issues for any problems blocking this criterion should be created and added to the subsequent sprint)