Re: Connecting cluster with different cookies

Would it be sufficient to be able to set the default cookie
for a node with e.g a new API function erlang:set_cookie(Cookie)?

/ Raimo Niskanen


On Mon, Aug 09, 2021 at 11:04:14PM +0300, Stanislav Ledenev wrote:
> Just  couple of thoughts:
> 
> Wouldn't the removal of erlang:set_cookie() have a huge impact on one of
> the cool
> features of Erlang - dynamic reconfiguration of the running system without
> stopping it?
> This function is the only dynamic way of setting a cookie. Two others are
> static, i.e.
> through the file and command line parameter.
> 
> And also for security demanding applications which work within insecure
> environments erlang:set_cookie() is a very useful function. I.e. read cookie
> (which changes in time) from a trusted storage and set it without stopping
> the
> whole system.
> 
> пн, 9 авг. 2021 г. в 20:08, Peter-Henry Mander <[email protected]>:
> 
> > Hi Raimo,
> >
> > Thanks for asking (-:
> >
> > We've found this erlang:set_cookie/2 function to be very handy in
> > Operations, to be able to launch a remote shell connecting to nodes in a
> > collection of clusters, each with different cookies. E.g. someone has
> > opened a quote in the to_erl shell, leaving the shell "broken" (please,
> > don't snigger...) in the heat of an incident.
> >
> > The use case is hardly critical, but the ability to swap cookies will be
> > missed.
> >
> > Pete.
> >
> >

-- 

/ Raimo Niskanen, Erlang/OTP, Ericsson AB