Options
error when creating cluster
I get this error when creating a new cluster. How to troubleshoot ?
The error seems to pop up regardless of the instance type selected.
Created iamRole: verticaeon1_1509581033274_203
Info Failed to provision AWS Cluster: [fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "msg": "Instance creation failed => Unsupported: EBS-optimized instances are not supported for your requested configuration. Please check the documentation for supported configurations."}]
Failed Provision step failed, please check log for details.
Info Rollback AWS resources.
I picked c3.4xlarge in us-east-2
0
Comments
According to my EC2 consultant:
The c3 is a legacy tier. New regions and some of the new data centers in legacy regions don’t support it anymore. What AZ are you in?
But the message is quite confusing sometimes with AWS provisioning. Even the right error message should be c3.4xlarge isn’t support in that availability zone, but AWS will complain no EBS optimization supported because of the validation process in instance launch is crewed.
The same thing we discovered with us-east-1c, we add the extra handling in playbook to avoid setting the EBS-optimzed flag, which will let AWS give the real reason – instance not supported.
Colin, you might want to see whether you can spin up those instances manually from the EC2 Dashboard - in other words, confirm that you can have three instances running concurrently with the MC on the 4th. This will eliminate the possibility that your AWS account has restrictions. For example, you may have run out of Elastic IPs or you might have exceeded the number of instances that your account can spin up.
Good suggestions, thanks Ben & Dave !
I retried the process with c4.4xl instead of c3. It fails with a different error message
Info Creating AWS Instances.
Info Created iamRole: verticaeon1_....
Info Created securityGroup: sg-...
Info Created privateIPs: [....]
Failed Provision step failed, please check log for details.
Info Rollback AWS Instances.
Info Rollback AWS resources.
but the ec2 hosts are created and started, but incompletely provisioned.
admintools is owned by root, and i did not dig deeper on what else is missing.
the zone is us-east-2.
How to troubleshoot?
Colin, let's set up 30 minutes to review what you are doing - I will email you directly to schedule.
The problem here was that a new AWS account had a low limit of 5 elastic IPs and once those were used up, the cluster creation process failed. The error was not visible in the MC screen as you see above, so that needs to be addressed. It would have saved some 'time' rather than digging in the MC logs. Thanks Dave for assisting.
Thanks for closing this issue.
I have entered a bug report regarding the obscurity of the error message.