fix(amazonq): fix for amazon q app initialization failure on sagemaker #7548
Conversation
|
| (Experiments.instance.get('amazonqLSP', true) || Auth.instance.isInternalAmazonUser()) && | ||
| (!isAmazonLinux2() || hasGlibcPatch()) | ||
| ) { | ||
| (isSageMaker() || !isAmazonLinux2() || (await hasGlibcPatch())) |
There was a problem hiding this comment.
it appears that the isAmazonLinux2 check is incorrectly flagging ubuntu containers under AL2 hosts. i do not think we need to special case sagemaker for now
There was a problem hiding this comment.
what is the exact setup for SageMaker AL2 instances?
| "@aws-sdk/smithy-client": "<3.731.0", | ||
| "@aws-sdk/util-arn-parser": "<3.731.0", | ||
| "@aws/mynah-ui": "^4.35.4", | ||
| "@aws/mynah-ui": "^4.35.6", |
There was a problem hiding this comment.
mynah update should not be in this focused change
There was a problem hiding this comment.
sorry..previous commit sneaked in...let me correct and raise a new one
|
What is the root cause of this issue? If SageMaker does not have such env var, why is it going to the GLIBC patch code path |
|
like richard said...that AL2 check |
| * @returns true if the current system is SageMaker(SMAI or SMUS) | ||
| */ | ||
| export function isSageMaker(appName: 'SMAI' | 'SMUS' = 'SMAI'): boolean { | ||
| // Check for SageMaker-specific environment variables first |
There was a problem hiding this comment.
FYI, the check for Unified Studio i.e process.env.SERVICE_NAME === sageMakerUnifiedStudio, is already handled in isSageMakerUnifiedStudio()
And with this logic, it will bypass the check for Sagemaker AI or Unified Studio cases right? Can we extract these env variable checks omitting this process.env.SERVICE_NAME === sageMakerUnifiedStudiointo a boolean and use it in case statements for better readability?
4 participants
Problem
GlibC path execution failing for sagemaker instances due to changes made to support AL2.
Solution
feature/xbranches will not be squash-merged at release time.