- Newest
- Most votes
- Most comments
Based on the error message you're encountering, it appears that your Knowledge Base role doesn't have proper access to the Amazon Titan embedding model. This is a common issue that can be resolved by following these steps:
The Root Cause
The main issue is likely that you haven't subscribed to the Amazon Titan Embedding model in the Bedrock Model Marketplace. Even if you have the correct IAM permissions, you need to explicitly subscribe to the model before you can use it.
Step-by-Step Solution
- Subscribe to the Titan Embedding Model:
- Go to the Amazon Bedrock console
- Navigate to "Model access" under "Bedrock configurations"
- Find "Amazon Titan Embed Text v2" in the list
- Click "Request model access" if it's not already enabled
- Complete the subscription process
- Verify IAM Role Permissions: Your IAM role for the Knowledge Base needs these permissions:
bedrock:InvokeModelfor the Titan embedding model- S3 access permissions for your data source
- S3Vectors or OpenSearch permissions depending on your vector store
- Create a Clean Knowledge Base Setup:
- Delete any existing problematic Knowledge Base
- Delete any auto-created roles that might be causing conflicts
- Create a new Knowledge Base with "Create and use a new service role"
- After creation, you can enhance the auto-created role with additional permissions if needed
- Required IAM Policy Elements: Your Knowledge Base role should have a policy that includes:
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"bedrock:InvokeModel"
],
"Resource": [
"arn:aws:bedrock:us-east-1::foundation-model/amazon.titan-embed-text-v2:0"
]
},
{
"Effect": "Allow",
"Action": [
"s3:ListBucket"
],
"Resource": [
"arn:aws:s3:::{YOUR_SOURCE_BUCKET_NAME}"
]
},
{
"Effect": "Allow",
"Action": [
"s3:GetObject"
],
"Resource": [
"arn:aws:s3:::{YOUR_SOURCE_BUCKET_NAME}/*"
]
}
]
}
- For OpenSearch Serverless Vector Store: Add the necessary OpenSearch permissions to your role if you're using OpenSearch Serverless as your vector store.
The most important step is subscribing to the model in the Bedrock Model Marketplace. Many users encounter this issue because they have the correct IAM permissions but haven't completed the model subscription process. After subscribing to the model and ensuring your IAM role has the proper permissions, you should be able to successfully sync your Knowledge Base.
Sources
Data sync failed error in Amazon Bedrock Knowledgebase | AWS re:Post
Resolve file sync failures with Amazon Bedrock Knowledge Bases | AWS re:Post
Building cost-effective RAG applications with Amazon Bedrock Knowledge Bases and Amazon S3 Vectors | Artificial Intelligence
answered a year ago
Hey,
Hope you're keeping well.
The sync error means the Knowledge Base’s service role doesn’t have permission to invoke the Titan embedding model. When you select “Create and use a new service role,” Bedrock creates an IAM role with S3 and OpenSearch access, but it won’t automatically include the bedrock:InvokeModel permission for that specific model ARN.
Go to IAM > Roles, open your Knowledge Base role, and attach an inline policy granting bedrock:InvokeModel on the Titan model ARN you’re using. For example:
{ "Version": "2012-10-17", "Statement": [ { "Effect": "Allow", "Action": "bedrock:InvokeModel", "Resource": "arn:aws:bedrock:us-east-1::foundation-model/amazon. Thanks and regards, Taz
answered a year ago
Relevant content
asked 2 years ago
- AWS OFFICIALUpdated 2 years ago
