https://pulumi.com logo
Join Slack
Powered by
Just started getting this error: ```error: could n...
# typescript
g
Just started getting this error:
Copy code
error: could not deserialize deployment: snapshot integrity failure; refusing to use it: resource urn:pulumi:stage1::infra::aws:lb:ApplicationLoadBalancer$aws:route53/record:Record::r53-resource-name dependency urn:pulumi:stage1::infra::aws:lb:ApplicationLoadBalancer$aws:lb/loadBalancer:LoadBalancer::lb-name refers to missing resource
I cannot simply 
rm
my stack as it's in use (and has been from some time). Importing the resource that is claimed to be missing is also failing.
c
What happened in the update immediately preceding the one where this started happening? And what happened exactly during this update? Was it interrupted somehow?
g
On the command preceding (before the error), I had tried a 
pulumi up --target ...
(specifying a DIFFERENT ALB), but I had neglected to properly quote the string. That's when I got errors related to that 
up
command. Now I get the error on almost every run.
c
Can you paste the output of that failed 
up --target
? You can get it from Pulumi Cloud by looking at the Updates tab and find that exact failed update. (Of course, please feel free to remove anything you think might be sensitive such as internal codenames/project names etc., if any.) You may DM it to me as well if you'd like. Also did you by any chance set 
deleteBeforeReplace
resource option on the load balancer resource?
g
Sorry, @clever-sunset-76585 - missed your comment earlier. No, I did not set 
deleteBeforeReplace
.
e
Faced similar Issue.
Steps Taken to Resolve Issue * Executed 
pulumi stack export --show-secrets --file stack.json
in order to study the stack file and noticed that search for 
"urn": "urn:pulumi:stack::infra::aws:lb:ApplicationLoadBalancer$aws:lb/loadBalancer:LoadBalancer::stack-internal"
, was returning nothing even though the resource existed in AWS. * Added the following snippet as indicated below to the stack.json (snippet generated from a similar resource):
Copy code
...
            {
                "urn": "urn:pulumi:stack::infra::aws:lb:ApplicationLoadBalancer$aws:lb/loadBalancer:LoadBalancer::stack-internal",
                "custom": true,
                "id": "arn:aws:elasticloadbalancing:us-west-2:************:loadbalancer/app/stack-internal/****************",
                "type": "aws:lb/loadBalancer:LoadBalancer",
                "parent": "urn:pulumi:stack::infra::aws:lb:ApplicationLoadBalancer::stack-internal",
                "dependencies": [
                    "urn:pulumi:stack::infra::aws:s3/bucketPolicy:BucketPolicy::stack-internal-accessLogsPolicy",
                    "urn:pulumi:stack::infra::awsx:x:ec2:SecurityGroup$aws:ec2/securityGroup:SecurityGroup::stack-internal"
                ],
                "provider": "urn:pulumi:stack::infra::pulumi:providers:aws::default_5_43_0::9009540a-405a-44d9-b0c1-c61d3996fcd2",
                "propertyDependencies": {
                    "accessLogs": [
                        "urn:pulumi:stack::infra::aws:s3/bucketPolicy:BucketPolicy::stack-internal-accessLogsPolicy"
                    ],
                    "securityGroups": [
                        "urn:pulumi:stack::infra::awsx:x:ec2:SecurityGroup$aws:ec2/securityGroup:SecurityGroup::stack-internal"
                    ],
                }
            },
            ...
* Tested out on local provider, this involves importing the export file created. More info, https://www.pulumi.com/docs/concepts/state/#migrating-between-state-backends. * Deleted the old stack on Pulumi Provider after successful testing on local provider.
Copy code
pulumi stack rm -f stack
* Initiated new stack and added configurations:
Copy code
pulumi stack init stack
pulumi config set --secret pagerduty:token ************
pulumi config set aws:region us-west-2

pulumi config
* Imported the edited export file.
Copy code
pulumi stack import --file stage.stack.json
* Refreshed the state to get all details of the missing resource.
Copy code
pulumi refresh
g
@clever-sunset-76585 I made some progress: I got past this issue by doing the following:
Copy code
# pulumi stack export --show-secrets --file stage.stack.json
I manually added the missing ALB resource into the JSON file, using one of the other ALBs as a template and substituting the correct values. Then (and this was the scary bit), removed the stack:
Copy code
# pulumi stack rm -f stage1
The re-initialized a new one:
Copy code
# pulumi stack init stage1
# pulumi config set --secret pagerduty:token ************
# pulumi config set aws:region us-west-2
# pulumi config
And, finally, imported the JSON:
Copy code
# pulumi stack import --file stage.stack.json
# pulumi refresh
This at least got past the error. But now, when I re-run my code (
pulumi up
) there's several resources that it wants to delete and/or otherwise modify.
c
Glad you past the issue, @gray-airplane-38353. You didn't have to remove the stack though. An import of the edited state file would have sufficed.
But now, when I re-run my code (
pulumi up
) there's several resources that it wants to delete and/or otherwise modify.
That's probably because, and I am assuming based on the snippet of commands you pasted above, you initialized a stack with a different name than the previous one. So the resource URNs are completely different. I am actually not sure how you were even able to import the checkpoint from a stack named 
stage
into a stack named 
stage1
. I would have thought that the CLI would complain that the stack names don't match. In any case, if you are still having a problem, you might try renaming your stack with 
pulumi stack rename ...
or force remove the stack and recreate a new stack yet again with the original name and re-import the state that you saved.
6 Views
Open in Slack
PreviousNext
Powered by