Implement auto-detection #41
Conversation
|
|
There was a problem hiding this comment.
Added a few comments.
You can also check the AWS PR for a reference for tests/javadocs, etc.
https://dzone.com/articles/the-stepdown-rule
| return DiscoveryStrategyLevel.CLOUD_VM; | ||
| } | ||
|
|
||
|
|
| } | ||
|
|
||
|
|
||
| private String readFileContents(String fileName) { |
There was a problem hiding this comment.
Consider changing this method and others to static
| } | ||
| } | ||
|
|
||
| private boolean googleInternalDnsConfigured() { |
There was a problem hiding this comment.
Please reorder the methods. They should follow Stepdown rule
There was a problem hiding this comment.
I did it following the rule.
src/main/java/com/hazelcast/gcp/GcpDiscoveryStrategyFactory.java
Outdated
Show resolved
Hide resolved
Co-authored-by: Rafał Leszko <rafal@hazelcast.com>
src/main/java/com/hazelcast/gcp/GcpDiscoveryStrategyFactory.java
Outdated
Show resolved
Hide resolved
| } | ||
|
|
||
| private static boolean metadataFlavorGoogle() { | ||
| return isEndpointAvailable("metadata.google.internal"); |
There was a problem hiding this comment.
This part does not work for me. I get the following exception.
com.hazelcast.gcp.RestClientException: Failure in executing REST call
at com.hazelcast.gcp.RestClient.call(RestClient.java:98)
at com.hazelcast.gcp.RestClient.get(RestClient.java:61)
at com.hazelcast.gcp.GcpDiscoveryStrategyFactory.isEndpointAvailable(GcpDiscoveryStrategyFactory.java:135)
at com.hazelcast.gcp.GcpDiscoveryStrategyFactory.metadataFlavorGoogle(GcpDiscoveryStrategyFactory.java:113)
at com.hazelcast.gcp.GcpDiscoveryStrategyFactory.isAutoDetectionApplicable(GcpDiscoveryStrategyFactory.java:78)
at com.hazelcast.spi.discovery.impl.DefaultDiscoveryService.detectDiscoveryStrategyFactory(DefaultDiscoveryService.java:207)
at com.hazelcast.spi.discovery.impl.DefaultDiscoveryService.loadDiscoveryStrategies(DefaultDiscoveryService.java:147)
at com.hazelcast.spi.discovery.impl.DefaultDiscoveryService.<init>(DefaultDiscoveryService.java:58)
at com.hazelcast.spi.discovery.impl.DefaultDiscoveryServiceProvider.newDiscoveryService(DefaultDiscoveryServiceProvider.java:29)
at com.hazelcast.instance.impl.Node.createDiscoveryService(Node.java:339)
at com.hazelcast.instance.impl.Node.<init>(Node.java:265)
at com.hazelcast.instance.impl.HazelcastInstanceImpl.createNode(HazelcastInstanceImpl.java:148)
at com.hazelcast.instance.impl.HazelcastInstanceImpl.<init>(HazelcastInstanceImpl.java:117)
at com.hazelcast.instance.impl.HazelcastInstanceFactory.constructHazelcastInstance(HazelcastInstanceFactory.java:211)
at com.hazelcast.instance.impl.HazelcastInstanceFactory.newHazelcastInstance(HazelcastInstanceFactory.java:190)
at com.hazelcast.instance.impl.HazelcastInstanceFactory.newHazelcastInstance(HazelcastInstanceFactory.java:128)
at com.hazelcast.core.Hazelcast.newHazelcastInstance(Hazelcast.java:91)
at com.hazelcast.core.server.HazelcastMemberStarter.main(HazelcastMemberStarter.java:47)
Caused by: java.net.MalformedURLException: no protocol: metadata.google.internal
at java.base/java.net.URL.<init>(URL.java:645)
at java.base/java.net.URL.<init>(URL.java:541)
at java.base/java.net.URL.<init>(URL.java:488)
at com.hazelcast.gcp.RestClient.call(RestClient.java:72)
Could you double check?
There was a problem hiding this comment.
Can you tell me what you ran for this to happen?
There was a problem hiding this comment.
Yeah, I just created a new GCP VM, copied there Hazelcast and Hazelcast Kubernetes plugin Jars and executed:
java -cp hazelcast-kubernetes.jar:hazelcast.jar com.hazelcast.core.server.HazelcastMemberStarter
Also you need to enabled debug logging or change the log level in the code.
There was a problem hiding this comment.
Do you mean gcp plug-ins?
There was a problem hiding this comment.
You were right, because I used curl while testing I forgot to add "http://" to the code.
|
|
||
| private static boolean serviceAccountAttached() { | ||
| try { | ||
| return isEndpointAvailable("metadata.google.internal/computeMetadata/v1/instance/service-accounts/default/"); |
There was a problem hiding this comment.
I think this part is not correct. What is someone does not use default service account? Or if someone does not have persmission to check the service accounts?
There was a problem hiding this comment.
If they have auto-configuration enabled, I assume they don't provide private-key-path. In that case only way to get token for compute api is through a service account. If they don't have any service accounts then GCP discovery won't work.
I don't understand what you mean by "if they don't have permissions to check service accounts"? As far as I know if they are on a GCP VM they can use instance metadata api without need for a permission, because it is only provided for that instance.
For the default service account, in "GcpMetadataApi" class we use default service account to get the token, so if they don't have a default one, discovery api wouldn't work nevertheless.
I am not confident about my findings though I am not familiar with gcp.
There was a problem hiding this comment.
If you define a different service account while creating a GCP VM, will it still work?
There was a problem hiding this comment.
Other than rafal? If the service account has permissions for reading project and region data, then I think it would work.
There was a problem hiding this comment.
If you checked that it works with other service account, then it's fine.
There was a problem hiding this comment.
I will check it again.
There was a problem hiding this comment.
It works correctly, I tried it with two different service accounts
No description provided.