Search code examples
bashapache-kafkachecksumsha512

Programmatically verify checksum using gpg format


I would like to programmatically validate the sha512 checksum of a Kafka binary. First I download the binary and the sha512 sum text file:

curl -fsSL -O \
    https://ftp.wayne.edu/apache/kafka/2.7.0/kafka_2.13-2.7.0.tgz
curl -fsSL -O \
    https://downloads.apache.org/kafka/2.7.0/kafka_2.13-2.7.0.tgz.sha512

I know by manual inspection that the checksum is ok:

$ cat -ne kafka_2.13-2.7.0.tgz.sha512 
     1  kafka_2.13-2.7.0.tgz: F3DD1FD8 8766D915 0D3D395B 285BFA75 F5B89A83 58223814$
     2                        90C8428E 6E568889 054DDB5F ADA1EB63 613A6441 989151BC$
     3                        7C7D6CDE 16A871C6 674B909C 4EDD4E28$

$ sha512sum kafka_2.13-2.7.0.tgz
f3dd1fd88766d9150d3d395b285bfa75f5b89a835822381490c8428e6e568889054ddb5fada1eb63613a6441989151bc7c7d6cde16a871c6674b909c4edd4e28  kafka_2.13-2.7.0.tgz

But shasum/sha512sum don't seem to like the format of the .512 file to do programatic validation (0 exit code on success, 1 on failure).

$ sha512sum --check kafka_2.13-2.7.0.tgz.sha512 
sha512sum: kafka_2.13-2.7.0.tgz.sha512: no properly formatted SHA512 checksum lines found

$ echo "$(cat kafka_2.13-2.7.0.tgz.sha512) kafka_2.13-2.7.0.tgz" \
    | sha512sum --check
sha512sum: 'standard input': no properly formatted SHA512 checksum lines found

What do I have wrong here? Is kafka_2.13-2.7.0.tgz.sha512 in an unconventional format or am I missing a command line flag?


Solution

  • Seems like kafka is using gpg --print-md sha512 https://github.com/apache/kafka/blob/trunk/release.py#L616

    Verification is done by diff then

    $ gpg --print-md SHA512 kafka_2.13-2.7.0.tgz | diff - kafka_2.13-2.7.0.tgz.sha512
    

    http://people.apache.org/~ke4qqq/ig/sect-source-verify.html