Getting ubuntu iso securely

J Fernyhough j.fernyhough at
Tue Sep 15 20:32:32 UTC 2015

OK - now you've lost me.

Earlier in the thread you were talking about PGP keys and web-of-trust, not
about verifying the integrity of a downloaded file.

You also mentioned a 10-line script to use as a downloader. Whoever is
downloading the file has to use some operating system to do so, whether
*nix or Windows. Any Linux or Mac install has (IIRC) sha256sum. Windows
users can use a GUI checksum utility.

If you're worried about users getting corrupt downloads, this is about user
education, not another technology solution (to a problem that's already
been solved). I wrote the Manjaro beginner's guide, and noone has
complained they don't understand how to check their downloaded installer
image. If there's one group who doesn't complain about documentation, it's

On 15 September 2015 at 20:53, Ryein Goddard <ryein.goddard at>

> If we are trying to target newbies that don't know what a sha256sum is
> then I highly doubt they will be running Ubuntu in order to run that
> command.
> Personally when I make an ubuntu ISO my CD burner program checks the value
> for it isn't an issue for me.  I am also not worried that it has
> been modified in transit, or my DNS requests have been spoofed.  I am more
> worried it hasn't been downloaded correctly.
> On Tue, Sep 15, 2015 at 12:48 PM, J Fernyhough <j.fernyhough at>
> wrote:
>> It's no more secure than running:
>> sha256sum -c ubuntu-installer.iso.shasum
>> or just:
>> sha256sum ubuntu-installer.iso
>> and manually checking the values match.
>> I'd even argue a script is less secure, as the user is running an
>> arbitrary script they've downloaded. It's also no more straightforward as
>> the user has to download and run the script. Whatever format the script is,
>> the user still has to set it as executable. By this point, reading a line
>> of instruction and running a single command is pretty trivial.
>> I understand what you're trying to do, I just think you're trying to
>> solve a problem that doesn't exist.
>> On 15 September 2015 at 20:40, Ryein Goddard <ryein.goddard at>
>> wrote:
>>> We are talking about a more secure method with a built in way to
>>> checksum that is easy for users not the Pentagon.
>>> On Tue, Sep 15, 2015 at 12:30 PM, J Fernyhough <j.fernyhough at>
>>> wrote:
>>>> An "open" script with an encrypted checksum? What's to stop someone
>>>> compromising this script during transport? You have recreated *exactly* the
>>>> same problem, just a level higher.
>>>> On 15 September 2015 at 20:27, Ryein Goddard <ryein.goddard at>
>>>> wrote:
>>>>> That part is easy because it could be a open script with probably less
>>>>> then 10 lines of code.
>>>>> On Tue, Sep 15, 2015 at 12:23 PM, J Fernyhough <j.fernyhough at
>>>>> > wrote:
>>>>>> And how would you know the Ubuntu-branded downloader is secure?
>>>>>> I think you're over-complicating things here. Anyone interested in
>>>>>> verifying a download is correct can verify the posted SHAsum, and anyone
>>>>>> really concerned could install from a netboot (mini.iso), check its seed
>>>>>> file, and download all packages from a known repo.
>>>>>> If you are concerned about an installer download becoming compromised
>>>>>> during transport then you should also be concerned about the apt transport
>>>>>> used - I'm assuming you set your deb sources to https? If not, then a
>>>>>> 'secure' installer image is moot.
>>>>>> J
>>>>>> On 15 September 2015 at 20:10, Ryein Goddard <ryein.goddard at
>>>>>> > wrote:
>>>>>>> You could add multiple sources that store an encrypted checksum and
>>>>>>> then reference that with an Ubuntu branded downloader.  That program would
>>>>>>> be pretty easy to make and it would abstract away all requirements for
>>>>>>> anything time consuming from the user.
>>>>>>> On Tue, Sep 15, 2015 at 3:53 AM, Ralf Mardorf <
>>>>>>> ralf.mardorf at> wrote:
>>>>>>>> On Mon, 14 Sep 2015 15:07:02 -0700, Ryein Goddard wrote:
>>>>>>>> >On Mon, Sep 14, 2015 at 10:32 AM, Ralf Mardorf wrote:
>>>>>>>> >> On Mon, 14 Sep 2015 16:19:36 +0000 (UTC), rajeev bhatta wrote:
>>>>>>>> >> >It is not time consuming.. just for the user experience..
>>>>>>>> >>
>>>>>>>> >> IMHO for averaged users it is time consuming. Even a power users
>>>>>>>> not
>>>>>>>> >> necessarily deals with the right people to get a key she or he
>>>>>>>> can
>>>>>>>> >> trust, that can be used to verify ownership of the particular
>>>>>>>> >> public Ubuntu key.
>>>>>>>> >>
>>>>>>>> >> I am a Linux power user and I don't own a key to verify the
>>>>>>>> >> particular public key, that belongs to the key, that was used to
>>>>>>>> >> sign the Ubuntu images.
>>>>>>>> >>
>>>>>>>> >> Please let me know, how I can get such a key, without spending
>>>>>>>> much
>>>>>>>> >> time ;).
>>>>>>>> >
>>>>>>>> >If a current method doesn't exist then maybe we can just create
>>>>>>>> one?
>>>>>>>> How will you make it less time consuming?
>>>>>>>> You need to meet other people in the real world, in addition you
>>>>>>>> need to know and trust those people and in addition they need to
>>>>>>>> trust a
>>>>>>>> chain of trusted keys, that confirms ownership of the public Ubuntu
>>>>>>>> key
>>>>>>>> in question.
>>>>>>>> This already is hard to realise for hardcore computer geeks and
>>>>>>>> completely illusorily for those who's centre of life isn't the
>>>>>>>> operating system of their computers or digital security.
>> --
>> Ubuntu-devel-discuss mailing list
>> Ubuntu-devel-discuss at
>> Modify settings or unsubscribe at:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Ubuntu-devel-discuss mailing list