Skip to content

aws s3 sync command is working according to filesize change instead of md5sum change for identical filename file #8377

Closed
@YoongLoong

Description

@YoongLoong

Describe the bug

Hi, i have a situation in my usage on aws s3 sync:
File A: abcdefg.txt (1234 Bytes) (md5sum: a954ea6b12db7d788350b4c008ee26fb)
File B: abcdefg.txt (1234 Bytes) (md5sum: bd84bb07f07a58e832bb3c72e3db4663)

After I aws s3 sync the File A to the s3 location and try to sync the File B to the same s3 location it will skipped. Sync is only taking care of "filesize change" only instead of md5sum change which defeat the purpose of "SYNC" and it will give the wrong impression that there is no change between File A and File B. Would this be fixed or taken out as it is confusing to the user that they have the identical files being "sync"?

Expected Behavior

It should "sync" File B and replace File A upon second aws s3 sync as it is the "different" file with the same filename (different md5sum).

Current Behavior

aws s3 sync command skipped the upload due to the "same" file being uploaded (same filesize different md5sum)

Reproduction Steps

echo 1 > A/test
echo 2 > B/test
aws s3 sync A/ s3://<S3_Bucket>/ --exclude "" --include "test" --no-progress
aws s3 sync B/ s3://<S3_Bucket>/ --exclude "
" --include "test" --no-progress

Possible Solution

ETag of "B/test" file should be calculated and compared with the s3://<S3_Bucket>/test to ensure it is the same file before skipped sync

Additional Information/Context

No response

CLI version used

2.13.6

Environment details (OS name and version, etc.)

ubuntu14.04

Metadata

Metadata

Assignees

Labels

duplicateThis issue is a duplicate.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions