Removing failed ECE deployments in Cloud UI

[Ryan_Downey]

ECE 2.2
Issue: I have two deployments showing in the Cloud UI that are stuck in the "Deployment Not Found" stage and I'm trying to clarify how to clear them out correctly.

From what I understand of the post linked below I should log in to one of my Coordinator nodes and run two commands since I dont have access to the API console.

1 `curl -XPOST -u 'admin:PASS' 'localhost:12400/api/v1/clusters/elasticsearch/ID/_shutdown?skip_snapshot=true'`
2 `curl -XDELETE -u 'admin:PASS' 'localhost:12400/api/v1/clusters/elasticsearch/ID`

The API reference table has some minor syntax changes but this seems like the general route that I need to take to remove them from the Cloud UI.

Failed attempts to create a deployment

ECE 2.0.0 Failed attempts to create a deployment result in a failed cluster showing in the list of deployments that cannot be removed. The deployment shows the following status: Deployment health issues: Some instances in the deployment are unhealthy The latest attempt to change the configuration failed When looking at the ECE Cloud UI, the node appears under the allocator with a Red X and the version is showing 'v'. However, when logging into the OS on the allocator the node is not showi…

[Alex_Piggott]

Yes - that sound right. Sorry about that!

Alex

[Ryan_Downey]

Alex, appreciate the quick response. With ECE are there any other steps that you need to take in order for these commands to work with a TLS enabled environment. Generally in our environment if I add -k or --insecure to the end of a command it goes through. As of right now I'm not getting that to work.

sudo curl -XPOST -u admin:<password from installation> "https://xx.xxx.xx.xxx:12443/api/v1/clusters/elasticsearch/<cluster id>/_shutdown?skip_snapshot=true" -k
{
  "errors": [{
    "code": "root.unauthorized.rbac",
    "message": "The supplied authentication is not authorized to access this resource. Reason: [Not Found]."
  }]

[Alex_Piggott]

Reason: [Not Found] implies the cluster id is wrong

If you curl api/v1/clusters/elasticsearch does the cluster id you have been using get returned?

If not, same question for api/v1/clusters/elasticsearch/_search?q=*

If it appears in the _search but not the GET then that is a known sync bug with an easy workaround (DELETE v1-elasticsearch-clusters-ece-region/elasticsearch-cluster/:id from the Adminconsole cluster, eg using that cluster's API console)

[Ryan_Downey]

Ok so I was able to get this all worked out which is appreciated! My Linux skills aren't very good so theres probably an easier way to do this but I got info with

curl -X POST -u admin:<password> "https://xx.xxx.xx.xx:12443/api/v1/clusters/elasticseach/_search?q=*" -k

Updated that with -o to send it to a file since I couldn't read all of the output

curl -X POST -u admin:<password> "https://xx.xxx.xx.xx:12443/api/v1/clusters/elasticseach/_search?q=*" -k -o cluster_post.txt

Ran the same thing with the GET

curl -X GET -u admin:<password> "https://xx.xx.xx.xx.xx:12443/api/v1/clusters/elasticsearh" -k -o cluster_get.txt

Checked the files, as you mentioned the cluster_id's appeared in the search but not the GET output

grep <cluster_id> cluster_post.txt
grep <cluster_id> cluster_get.txt

Cloud UI --> admin-console-elasticsearch --> Elasticsearch -> API Console --> DELETE v1-elasticsearch-clusters-ece-region/elasticsearch-cluster/<cluster_id>

Worked perfectly. Thanks again for the quick help @Alex_Piggott!!!