removing state.address

John Arbash Meinel john at arbash-meinel.com
Sun Aug 25 12:55:58 UTC 2013


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2013-08-22 23:00, Nate Finch wrote:
> The signature of state.address is identical to instance.Address.
> It looks like state.address only exists to convert to and from 
> instance.Address for use in state.machineDoc. My guess is that this
> is some duplication that slipped in during the addressability
> changes.
> 
> Simply changing machineDoc to have []instance.Address instead of 
> []address (and removing the conversion code) compiles and tests
> fine. Any objections to making this change? (interesting note,
> state.address has bson struct tags on a couple fields, but the
> tests didn't complain when they were missing... that's probably a
> hole in the tests. I'd probably move over the bson tags to
> instance.Address, unless someone tells me they're not needed /
> wanted).
> 
> -Nate

My only concern is if this is intended to be used in our High
Availability proposal. (Record the addresses of all state machines in
the db so that we can easily find other servers from a single server
when HA is working.)

I'm pretty hesitant to change the content of the DB unless we are
confident about the use cases of what it was there for. Though I'm
tempted to "if it isn't tested, its broken anyway, so strip out
unnecessary stuff unti we find we actually need it".


Maybe we can get feedback from William or Roger about this one.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (Cygwin)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlIZ/t4ACgkQJdeBCYSNAAN8MwCgk7Bs4T9cDNz+F/cOj5aUfwMl
WAoAoLZJJzlUQUpJAb5K0WGOgdtg+tPg
=tcF2
-----END PGP SIGNATURE-----



More information about the Juju-dev mailing list