This release includes a new feature to allow Score servers to handle downloading files from an S3 backend when the .meta file describing a file upload is not found. In normal operation, this feature will have no impact on Score servers. If for any reason the Score-generated .meta files cannot be found, then the download service will calculate the available parts and file description from the S3 object metadata. In particular, the object MD5 can now be read either from the built-in S3 API's ContentMD5 property, or in a custom user Metadata field. This provides more flexibility for Score to serve files that were not originally uploaded to the S3 server through the Score upload CLI. This also makes it download from S3 objects that were transferred from an Azure instance to S3.
One additional change is included that removes Score's dependency on ICGC specific Java packages. The related code has been added to the Score Core package.
Description
This release includes a new feature to allow Score servers to handle downloading files from an S3 backend when the .meta file describing a file upload is not found. In normal operation, this feature will have no impact on Score servers. If for any reason the Score-generated .meta files cannot be found, then the download service will calculate the available parts and file description from the S3 object metadata. In particular, the object MD5 can now be read either from the built-in S3 API's ContentMD5 property, or in a custom user Metadata field. This provides more flexibility for Score to serve files that were not originally uploaded to the S3 server through the Score upload CLI. This also makes it download from S3 objects that were transferred from an Azure instance to S3.
One additional change is included that removes Score's dependency on ICGC specific Java packages. The related code has been added to the Score Core package.
Included Changes
402
413
414