[HADOOP-18743] convert declarations of AWS v1 SDK EnvironmentVariableCredentialsProvider to v2 version

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Assignee: Unassigned
• Reporter: Steve Loughran
• Created: 17/May/23 13:44
• Updated: 27/Jul/23 14:39
• Resolved: 27/Jul/23 14:39

Description

As I play with the v2 sdk I've cut the v1 sdk from the tools/lib and now I'm getting stack traces about missing class EnvironmentVariableCredentialsProvider

java.io.IOException: From option fs.s3a.aws.credentials.provider java.lang.ClassNotFoundException: Class com.amazonaws.auth.EnvironmentVariableCredentialsProvider not found
        at org.apache.hadoop.fs.s3a.auth.AwsCredentialListProvider.loadAWSProviderClasses(AwsCredentialListProvider.java:128)
        at org.apache.hadoop.fs.s3a.auth.AwsCredentialListProvider.buildAWSProviderList(AwsCredentialListProvider.java:167)
        at org.apache.hadoop.fs.s3a.auth.AwsCredentialListProvider.createAWSCredentialProviderSet(AwsCredentialListProvider.java:102)
        at org.apache.hadoop.fs.s3a.S3AFileSystem.bindAWSClient(S3AFileSystem.java:945)
        at org.apache.hadoop.fs.s3a.S3AFileSystem.initialize(S3AFileSystem.java:597)
        at org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:3595)
        at org.apache.hadoop.fs.FileSystem.access$300(FileSystem.java:171)
        at org.apache.hadoop.fs.FileSystem$Cache.getInternal(FileSystem.java:3696)
        at org.apache.hadoop.fs.FileSystem$Cache.getUnique(FileSystem.java:3653)
        at org.apache.hadoop.fs.FileSystem.newInstance(FileSystem.java:608)
        at org.apache.hadoop.fs.store.diag.StoreDiag.executeFileSystemOperations(StoreDiag.java:753)
 
this provider is listed in the fs.s3a.aws.credentials.provider chain

<property>
<name>fs.s3a.aws.credentials.provider</name>
<value>
org.apache.hadoop.fs.s3a.TemporaryAWSCredentialsProvider,
org.apache.hadoop.fs.s3a.SimpleAWSCredentialsProvider,
com.amazonaws.auth.EnvironmentVariableCredentialsProvider,
org.apache.hadoop.fs.s3a.auth.IAMInstanceCredentialsProvider
</value>
...

Ideally we should remove all v1 dependencies, explicitly (here) and implicitly.

maybe for the env vars, we should consider adding our own env var provider which could go into branch-3.3 now as a single patch, which can then be cherrypicked by anyone who wants it in older releases; the v2 version will work with the new api

Issue Links

• Is contained by: HADOOP-18820 AWS SDK v2: make the v1 bridging support optional