Support Questions

Find answers, ask questions, and share your expertise

can't update existing cluster due to an instance image ami was not available

Contributor

Hi All,

 

I have created a cluster with the help of Cloudera Director 2.8 and in AWS , which was running successfully

 

Now I am trying to add few nodes to my existing cluster., when I updated the number through Web UI and continued to add I've got following error

 

[2020-06-05 13:37:10.682 -0400] ERROR [qtp1536478396-14] 47e9dd42-f37c-4c61-959a-49759e97c6f9 PUT /api/v12/environments/test_ha_cluster%20Environment/deploymen
ts/test_ha_cluster%20Deployment/clusters/test_ha_cluster - - c.c.l.a.c.ValidationExceptionHandler: Found validation issues
com.cloudera.launchpad.api.common.ClustersResource$InvalidClusterTemplateException: Validation errors:
ErrorInfo{code=PROVIDER_EXCEPTION, properties={message=Image (AMI) ID not found: ami-0b3fb240e8e9446a4}, causes=[]}
ErrorInfo{code=PROVIDER_EXCEPTION, properties={message=Image (AMI) ID not found: ami-0cdc5bfa27d60c283}, causes=[]}Conditions:
ValidationExceptionCondition{scope='aws.com.cloudera.director.aws.ec2.EC2Provider.template', key='Optional.of(image)', exceptionInfo='ErrorInfo{code=PROVIDER_E
XCEPTION, properties={message=Image (AMI) ID not found: ami-0b3fb240e8e9446a4}, causes=[]}'}
ValidationExceptionCondition{scope='aws.com.cloudera.director.aws.ec2.EC2Provider.template', key='Optional.of(image)', exceptionInfo='ErrorInfo{code=PROVIDER_E
XCEPTION, properties={message=Image (AMI) ID not found: ami-0cdc5bfa27d60c283}, causes=[]}'}

 

From the error I found that 2 AMI's were not found, and those AMI were deleted since we applied patches/updates ( After every patch we generate new AMI and retain only last 2 AMIs ).

 

After knowing that my instance template was refering to old AMIs I tried to update Template with cloudera director sdk ( python ) and now I am getting error as Unsupported cluster update

 

[2020-06-06 22:46:13.134 -0400] ERROR [qtp1536478396-11853] 21c7fc58-6642-4f3d-a7f1-97d84cee03f8 PUT /api/v12/environments/test_ha_cluster%20Environment/deployments/test_ha_cluster%20Deployment/clusters/test_ha_cluster - - c.c.l.api.common.ClustersResource: Unsupported cluster update
java.lang.IllegalArgumentException: Changes on the existing virtual instances are not supported: VirtualInstance{id='d73ad9d3-c9eb-492d-9c14-669a14a2795e', template=InstanceTemplate{name='zookeeper', type='m5.large', image='ami-0a82cefb61fce6193', rackId='/default-1', bootstrapScriptsArePresent=true, config={subnetId=subnet-003b5799ff0b54207, iamProfileName=test-east-zookeeper-role, ebsOptimized=false, rootVolumeSizeGB=100, ebsVolumeCount=0, enableEbsEncryption=false, rootVolumeType=gp2, ebsVolumeSizeGiB=100, useSpotInstances=false, ebsVolumeType=gp2, securityGroupsIds=sg-0b108e95420ad3914, spotBidUSDPerHr=0.1}, tags={}, normalizeInstance=true, normalizationConfig=Optional.absent(), sshUsername=Optional.absent(), sshHostKeyRetrievalType=NONE}} -> VirtualInstance{id='d73ad9d3-c9eb-492d-9c14-669a14a2795e', template=InstanceTemplate{name='zookeeper', type='m5.large', image='ami-026c0c75492919552', rackId='/default-1', bootstrapScriptsArePresent=true, config={subnetId=subnet-003b5799ff0b54207, iamProfileName=test-east-zookeeper-role, ebsOptimized=false, rootVolumeSizeGB=100, ebsVolumeCount=0, enableEbsEncryption=false, rootVolumeType=gp2, ebsVolumeSizeGiB=100, useSpotInstances=false, ebsVolumeType=gp2, securityGroupsIds=sg-0b108e95420ad3914, spotBidUSDPerHr=0.1}, tags={}, normalizeInstance=true, normalizationConfig=Optional.absent(), sshUsername=Optional.absent(), sshHostKeyRetrievalType=NONE}}

 

Is there a way I can update existing cluster instance ( /virtual instance groups ) image ( AMI )? or skipping this validation step and proceed with cluster update of adding new nodes?

 

@dhan , or @bhavanki Could you be able to help?

 

Any help in this is greatly appreciated

 

Thanks in advance

2 REPLIES 2

may be it can help: 

AMIs are specific to regions

(for ex: when a AMI is copied to another region, it will have a different id)

 

Contributor

Thanks @rachid-berkane  for quick response.

 

Please note that we're using only one region and in that only one AZ ( I should've mentioned in my post ). I think region is not the problem since we know that AMI's got deleted (in our process).

Take a Tour of the Community
Don't have an account?
Your experience may be limited. Sign in to explore more.