On 26/06/14 15:10 +0000, Kurt Griffiths wrote:
Crew, I’d like to propose the following:

1. Decouple pool management from data storage (two separate drivers)
2. Keep pool management driver for sqla, but drop the sqla data storage driver
3. Provide a non-AGPL alternative to MongoDB that has feature parity and is at
   least as performant

Decoupling will make configuration less confusing, while allowing us to
maintain drivers separately and give us the flexibility to choose the best tool
for the job (BTFJ). Once that work is done, we can drop support for sqla  as a
message store backend, since it isn’t a viable non-AGPL alternative to MongoDB.
Instead, we can look into some other backends that offer a good mix of
durability and performance.
What does everything think about this strategy?

+1 for the above. Lets make it a blueprint and start working on it.
The sooner we remove sqlalchemy as a message store, the less confused
people will be.

As for the viable non-AGPL alternative, I'll start a new thread with a
list of possible options.

Flavio

--
@flaper87
Flavio Percoco

Attachment: pgpgDC1EqRwtD.pgp
Description: PGP signature

_______________________________________________
OpenStack-dev mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to