Deployment

Deploy dotcom using the dotcom:frontend-all project in riff-raff. This will autodeploy in PROD- and it's usually what you want to do when using CODE.

What to do when a deployment breaks

com.amazonaws.AmazonServiceException Status Code: 400, AWS Service:
AmazonAutoScaling, AWS Request ID: e7abc4ec-8518-11e3-b64f-13b3661c52c1,
AWS Error Code: ValidationError, AWS Error Message: Currently,
desiredSize equals minSize (1). Terminating instance without replacement
will violate group's min size constraint. Either set
shouldDecrementDesiredCapacity flag to false or lower group's min size.
Show exception details

• The deployment process is attempting to scale down, but this would leave the autoscaling group empty.
• You need to re-deploy this project, Eg. gu deploy --code -n <project>

java.lang.RuntimeException Autoscaling group does not have the capacity
to deploy current max = 2 - desired max = 4 Show exception details

• A previous deployment or autoscale action has failed to scale down the infrastructure.
• You need to scale down the autoscale group so it's got room to add your servers, Eg, gu groups update <group> 3 3 6

com.amazonaws.AmazonServiceException Status Code: 400, AWS Service:
AmazonElasticLoadBalancing, AWS Request ID:
fad7ad53-8516-11e3-a6c5-937056a56545,

• A communication problem between the deployment and load balancer.
• Check the infrastructure is in a good state then try again.

AWS Error Code: Throttling, AWS Error Message: Rate exceeded Show
exception details

• We've exceeded our AWS rate-limit.
• Wait 5 minutes and re-deploy.

java.lang.RuntimeException Check failed to pass within 1200000
milliseconds (tried 41 times) - aborting Show exception details

• A box (or several boxes) has failed to launch and pass it's healthcheck meaning the healthcheck URL is not passing.
• Potentially you've released something that means the box can't be started or added to the load balancers.
• Does the healthcheck look like it is working? Eg, find the server, gu ec2 list then curl it's heathcheck, <host>:18080/management/healthcheck
• Log on to the box and have a look in the boot-up logs for errors and warnings.
• Ultimately, you'll probably need to check the size of the autoscale group and redeploy, but don't do this without understanding why it failed.

deployment.DeployStoppedException Deploy has been stopped by Matt
Chadburn Show exception details

• Someone (me in this case!) stopped your build manually.
• If that person isn't you then go and talk to them about why they stopped it as there may be a problem.

com.amazonaws.services.s3.model.AmazonS3Exception Status Code: 400, AWS
Service: Amazon S3, AWS Request ID: 292408DD7D4DBAD1, AWS Error Code:
RequestTimeout, AWS Error Message: Your socket connection to the server
was not read from or written to within the timeout period. Idle
connections will be closed. Show exception details

• The build artefacts could not be downloaded from the S3.
• The problem is probably temporary, so it's safe to try deploying again.