On 17.04.2017 22:59, sri wrote:
> Allright, i'll try to take the topic a little more serious. So straight
> to the point, there is no chance we would accept a change for a security
> feature that results in new attack vectors.

Very sensible IMO.

When I decided on a method for the secret(s), I knew I wanted to rotate
them, but also keep them on external storage (the DB, as I use one
anyway) so they can be reloaded on an application restart.

I found that rotating secrets in the DB from the Mojolicious
application, which was so simple in the development instance, turned out
to be tricky in the multi-process test and live environments. So I chose
to use an external process that regularly updates the list of secrets in
the DB, while the application processes check there for a new secret
regularly (and simultaneously). The external process takes (longish)
secrets from /dev/urandom, which should be secure enough.

Regards, Juergen.

-- 
Jürgen Nickelsen <[email protected]>
Freie Universität Berlin, ZEDAT/FUDIS https://zedat.fu-berlin.de/FUDIS
Jabber [email protected]; Tel +49.30.838-50740 Fax -450740

-- 
You received this message because you are subscribed to the Google Groups 
"Mojolicious" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/mojolicious.
For more options, visit https://groups.google.com/d/optout.

Reply via email to