69

I keep on finding myself wanting to download and check the integrity of the download immediately in a script, but I haven't been able to find the right incantation of sha256sum.

MY_SHA256=e147f0392686c40cfd7d5e6f332c6ee74c4eab4d24e2694b3b0a0c037bf51dc5
sha256sum some_binary | sha256sum --check ${MY_SHA256}

How can I take the sha256sum of a new file and compare it with a known hash immediately?

Improve this question
1

10 Answers 10

Reset to default
47

I have downloaded an archive file and an accompanying checksum file. Here is how I verify that the hash of the downloaded archive matches the hash from the downloaded checksum file:

echo "$(cat archive.tar.gz.sha256) archive.tar.gz" | sha256sum --check --status

The --status flag prevents all stdout output (more effective than --quiet). I then need to rely on the return code to determine if they matched, which is what I want anyway since I'm going to be using this in a script.

Improve this answer
4
  • 3
    Well, in my case, since file sha512 already contains the filename, so what i should do is echo "$(cat filename.sha512)" | sha512sum --check --status and immediately check return value using $?
    – fa wildchild
    May 18, 2020 at 15:03
  • 5
    It’s almost never useful to do echo "$(some-command)".  Just run the command; for example, cat filename.sha512 | sha512sum --check --status.  But that’s a UUOC; do sha512sum --check --status < filename.sha512 instead, or even sha512sum --check --status filename.sha512 (without the <).  Which, in fact, is fairly straightforward and obvious from the documentation.  … (Cont’d)
    – Scott - Слава Україні
    Feb 18, 2021 at 19:42
  • P.S. That’s fine if you’re writing a script, although you shouldn’t need to reference $? explicitly. But, if you’re doing this manually, and you’re doing echo $? and looking at the output, you might as well leave off the --status and let sha512sum tell you the result.
    – Scott - Слава Україні
    Feb 18, 2021 at 19:42
  • 4
    You don't need to do this. A much simpler way to do this is to sha256sum --check CHECKSUM_FILE. See my answer for details.
    – Debajit
    Aug 16, 2022 at 22:55
31

Example:

 echo "67574ee0039eaf4043a237e7c4b0eb432ca07ebf9c7b2dd0667e83bc3900b2cf kali-linux-2019.2-amd64.iso" | sha256sum -c

In case you have the sha256sum file, you can directly use it:

sha256sum -c "kali-linux-2019.2-amd64.iso.txt.sha256sum"

Explanation:

In the above example, you have

echo "<known SHA 256 sum of the file> <name of the file>" | sha256sum -c

sha256sum -c option can either read the SHA256 sum from a sha256sum file or from STDIN. In case you don't have the sha256sum file, then using the echo command you can provide the same details contained in a sha256sum file.

In case you have the sha256sum file, you can directly use it:

sha256sum -c "<sha256sum file name>"

Note:

Alternatively, you can use shasum -a 256 instead of sha256sum where -a specifies the algorithm to be used.

Improve this answer
30

You can see that sha256sum --check takes the output of a previous (regular) sha256sum run: it takes hashes and filenames via stdin, and compares them against actual files.

So the obvious thing to do is to manually give it the output in the format it wants:

$ echo "da39a3ee5e6b4b0d3255bfef95601890afd80709  motd" | sha1sum --check
motd: OK
Improve this answer
1
  • 1
    It seems to be taking the path to the file given in the std input and performing another sha256sum on that.
    – tarabyte
    Apr 10, 2018 at 20:47
4

All about checksums, including basic info. and usage

TLDR;

# 1. Check to see if file "filename" has this expected hash:
# `expected_checksum_hash`
echo "expected_checksum_hash filename" | sha256sum --check 

# 2. Check to see if these two files ("path/to/file1" and "path/to/file2")
# have the same checksum hash
echo "$(sha256sum "path/to/file1" | gawk '{ print $1 }') path/to/file2" \
| sha256sum --check

# OR (same as #2 just above)
file1_hash="$(sha256sum "path/to/file1" | gawk '{ print $1 }')" \
&& echo "$file1_hash path/to/file2" | sha256sum --check

DETAILS:

1. Background info

Note: you can use sha256sum or sha512sum in any of the examples below. These are the recommended and most-robust cryptographic checksums, with sha512sum, of course, being stronger.

There is also md5sum, but it isn't as robust, but is still commonly used for data integrity checks. Whenever possible, I recommend you use sha256sum or sha512sum instead. Wikipedia states that md5sum is still good for data integrity checks, but is "no longer deemed secure" and shouldn't be used for cryptographic purposes. So, just use sha256sum or sha512sum above, instead.

There are even more, however. Here is a list of the various checksum program you can technically use in any of the examples below:

sha1sum
sha224sum
sha256sum
sha384sum
sha512sum
shasum  # general-purpose tool, requires specifying the algorithm
md5sum

2. Get the checksum of a file:

sha256sum path/to/any/file

Example:

$ sha256sum FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz
6b579bd4ecdf86f7e70a009886c511da0b5085b831b0d6afc42442cabc249b90  FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz

Notice that the output of the sha256sum command is the numerical checksum hash followed by the file name this checksum corresponds to. You can store this checksum into a file named sha256sum.txt like this:

sha256sum path/to/file > sha256sum.txt

3. Compare the checksum of a file against a previously-stored or already-known checksum:

Now, assuming you want to check the integrity of the file against this known checksum in that file, you can test the file again like this:

# This causes the program to re-do the checksum of the file specified inside
# sha256sum.txt, and then compare it to the checksum in that same file. If they
# (the re-calculated checksum and the previously-stored checksum) match, it will 
# output the name of the file followed by "OK". 
sha256sum --check sha256sum.txt

Example:

$ sha256sum --check sha256sum.txt
FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz: OK

You can also manually pipe these things (the expected checksum hash and filename) to the checksum program, like this. This is really useful for when you need to check a downloaded file against a known checksum published online where you downloaded it. This way you can check for data integrity to ensure the downloaded file was downloaded successfully.

# 1. pipe to the checksum program directly
echo "expected_checksum_hash filename" | sha256sum --check 

# 2. OR, manually create the checksum file, and *then* run it on that file
# as done above
echo "expected_checksum_hash filename" > sha256sum.txt 
sha256sum --check sha256sum.txt  # same as previously done above

Example of option 1 just above:

$ echo "6b579bd4ecdf86f7e70a009886c511da0b5085b831b0d6afc42442cabc249b90 \
> FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz" | sha256sum --check
FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz: OK

4. To compare the checksum of file1 to file2:

Sometimes you have two downloaded files, or two copies of what you think are the same file, and you just want to ensure they are in fact the same (or different). Building on the information above, there are a few ways to do this.

  1. Manually check the checksum of each file, manually looking at the hashes to ensure they match:

    sha256sum 'path/to/file1'
sha256sum 'path/to/file2'
# now visually inspect both hashes

  2. OR [RECOMMENDED] automatically test file1 against file2:

    # Do some trickery to compare the hash of file1 agains the hash of file2. 
# Effectively, what we have done is this: 
# `echo "checksum_hash_from_file1 path/to/file2" | sha256sum --check`
# This therefore is checking to see if the hash from file1 matches the hash
# from file2. 
echo "$(sha256sum "path/to/file1" | gawk '{ print $1 }') path/to/file2" \
| sha256sum --check

# OR (same as just above)
file1_hash="$(sha256sum "path/to/file1" | gawk '{ print $1 }')" \
&& echo "$file1_hash path/to/file2" | sha256sum --check

    The way this works is that first it checks the checksum of file1, piping the output (hash and filename) to gawk, which is the GNU version of awk, which is a pattern-matching and text processing language. The gawk '{ print $1 }' command simply says to strip the first space-separated text field (indicated by $1), and retain it only. That's the checksum hash from file1. Then, we append the path/to/file2 and pipe this whole thing to be checked, as done previously above.

    In effect, we are tricking the checksum program into thinking we have a previously-obtained hash from file2, and we'd like to check it against a newly-calculated hash from file2. Since we used the hash from file1, however, but the filename of file2, we know that if it passes it is really saying file1 and file2 have the same hash, and are therefore identical files.

    Example:

    # technique 1
$ echo "$(sha256sum "FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz" \
| gawk '{ print $1 }') FoxitReader.enu.setup.2.4.4.0911_NEW.x64.run.tar.gz" \
| sha256sum --check
FoxitReader.enu.setup.2.4.4.0911_NEW.x64.run.tar.gz: OK

# technique 2
$ file1_hash="$(sha256sum "FoxitReader.enu.setup.2.4.4.0911.x64.run.tar.gz" \
| gawk '{ print $1 }')" \
&& echo "$file1_hash FoxitReader.enu.setup.2.4.4.0911_NEW.x64.run.tar.gz" \
| sha256sum --check

FoxitReader.enu.setup.2.4.4.0911_NEW.x64.run.tar.gz: OK
Improve this answer
2
  • To compare if two sums are the is it not enough to do only diff ... ?
    – user2820579
    Apr 7, 2022 at 9:05
  • @user2820579, diff is designed to "compare files line by line", meaning it is intended for text files, and it will show which lines differ, and how. sha256sum, on the other hand, is intended to compare any file, binary files in particular, and it will not show the difference between the files line-by-line, since the concept of "lines" doesn't exist in binary files. Instead, each binary file gets a unique "hash", which is like a signature for the file which is probabilistically unique for a given sequence of bytes in the file.
    – Gabriel Staples
    Apr 7, 2022 at 20:52
2

If you have the checksum file, you can simply do a sha256sum --check. For example:

% ls
elasticsearch-7.9.3-darwin-x86_64.tar.gz           # File to check
elasticsearch-7.9.3-darwin-x86_64.tar.gz.sha512    # Checksum file


% sha512sum --check elasticsearch-7.9.3-darwin-x86_64.tar.gz.sha512
elasticsearch-7.9.3-darwin-x86_64.tar.gz: OK
Improve this answer
1

The easiest way, the file with the sha512sums (in my case, "clear-36010-live-desktop.iso-sha512sums") and the file to verify in the same directory, and run

cat clear-36010-live-desktop.iso-sha512sums | sha512sum --check

Note: the content of the sha512sums file is both the value and the filename.

cat clear-36010-live-desktop.iso-sha512sums
8a47e79fc14721b5d0084f03c50ae66d6d7f7eeebb8d80a1dd168b2a2d5e31cb11efd69e202230b85914d159e0a0fbe20d57df08b825fa5aec0eea62fc1d2e64  clear-36010-live-desktop.iso
Improve this answer
1

not exactly answering the question but I made a function shacheck which compares 2 files using their sha256 hash and reports if files are identical or not (used to check file copy went well for example). I put this function in .bashrc and checked, it works. Any comments appreciated.

Listing:

#GOAL: check 2 files are identical using sha256
#$1=1st filename ; $2=2nd filename
shacheck () {

if [ "$#" -ne 2 ]; then 
    echo "number of arguments passed is $#";
    echo 'This function requires exactly 2 parameters, file 1 and file 2 and check they are the same using sha256 checksum';
    return -1;
else
#   number of arguments passed is 2

if [ -f $1 ]; then
    ARG1ISFILE=1
else
    ARG1ISFILE=0 
fi
if [ -f $2 ]; then
    ARG2ISFILE=1
else
    ARG2ISFILE=0
fi

#echo "\$ARG1ISFILE" $ARG1ISFILE
#echo "\$ARG2ISFILE" $ARG2ISFILE

TWOARGSAREFILE=$(($ARG1ISFILE + $ARG2ISFILE))
#echo "\$TWOARGSAREFILE" $TWOARGSAREFILE

if [ $TWOARGSAREFILE != 2 ]; then
    echo "both arguments have to be plain files"
    return -1;
else
    #echo "both arguments are plain files"
    CHECKSUMFILE1=$(sha256sum $1 | awk '{print $1}')
    CHECKSUMFILE2=$(sha256sum $2 | awk '{print $1}')

        if [ "$CHECKSUMFILE1" = "$CHECKSUMFILE2" ]; then
        #    echo "files are identical. \$CHECKSUMFILE1 = $CHECKSUMFILE1 \$CHECKSUMFILE2 = $CHECKSUMFILE2"

        #what I want as printed format:
        # files are identical. new line
        # CHECKSUM FILE 1 = $CHECKSUMFILE1 new line
        # equals new line
        # CHECKSUM FILE 2 = $CHECKSUMFILE2 new line
            echo -e "files are identical.\n"
            echo -e "CHECKSUM FILE 1 = $CHECKSUMFILE1 \n"
            echo -e "equals \n"
            echo -e "CHECKSUM FILE 2 = $CHECKSUMFILE2 \n"

            else

                echo -e "files are different.\n"
            echo -e "CHECKSUM FILE 1 = $CHECKSUMFILE1 \n"
            echo -e "is different than \n"
            echo -e "CHECKSUM FILE 2 = $CHECKSUMFILE2 \n"
    
        fi
    fi
fi
}
Improve this answer
0

I had an scenario where the sha256 files did not have the file name in it, no properly formatted SHA256 checksum lines found, as an example:

cat file.sha256
5b2266edc48f958687ab0f597f62dde1c6f01693d988e272411d081292991476

so I had to get each file with .sha256 read it to get the sha and then format the input to sha256sum command, like this:

for file in *.sha256; do echo $(cat $file) $(basename $file .sha256) | sha256sum --check; done
Improve this answer
0

Many software projects provide a checksum file some-release.zip.sha256 in addition to the downloadable file some-release.zip.

On one hand, the filename-checksum entries in .sha256 files use relative paths. On the other hand, the sha256sum command takes the current working directory from which it is called into account.

Assuming some-release.zip.sha256 is located within directory /tmp/, the following example invocation will work regardless of the current working directory, and will keep the current working directory after execution as it was before execution:

(cd /tmp/ && sha256sum --check some-release.zip.sha256)

On successful checksum verification, it will print

some-release.zip: OK

to stdout and exit with exit status 0.

Improve this answer
0

For OSXUsers (MacUsers) it´s possible to use:

echo "CHECKSUM" | shasum -a 256 PATHTOFILE | shasum -c

It´s no script, but it works for OSX ;-). Maybe someone can use it.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .