The Java example at https://github.com/awsdocs/amazon-glacier-developer-guide/blob/master/doc_source/checksum-calculations.md contains a bug in getChunkSHA256Hashes. The offset variable is misused when reading from the fileStream into the buff buffer. As soon as you try to hash a file larger than one chunk size (1MiB) it will thrown an IndexOutOfBounds exception:

Exception in thread "main" java.lang.IndexOutOfBoundsException
        at java.base/java.io.FileInputStream.readBytes(Native Method)
        at java.base/java.io.FileInputStream.read(FileInputStream.java:279)
        at TreeHashExample.getChunkSHA256Hashes(TreeHashExample.java:97)
        at TreeHashExample.computeSHA256TreeHash(TreeHashExample.java:56)
        at TreeHashExample.main(TreeHashExample.java:28)

Removing the offset variable and changing the while-loop to:

            while ((bytesRead = fileStream.readNBytes(buff, 0, ONE_MB)) > 0) {
                md.reset();
                md.update(buff, 0, bytesRead);
                chunkSHA256Hashes[idx++] = md.digest();
            }

will make it work correctly.

Hi Randy..this is a test.

On this topi.. i see a spelling error.

Glacier topic blah ..has this test issue. fix it.

Correct me if I'm wrong, but in this doc we mention bulk retrieval cost the least but we don't say how to calculate.

I did some search and found it's free:

https://aws.amazon.com/s3/glacier/pricing/
this was also mentioned in https://aws.amazon.com/about-aws/whats-new/2021/11/amazon-s3-glacier-storage-class-amazon-s3-glacier-flexible-retrieval/

If it is true, I think it'd be more user-friendly for the doc to just say so("cost the least" is technically true but ambiguous ).