Tobe,

The BP for the feature that added replica_count is here - 
https://github.com/openstack/trove-specs/blob/master/specs/kilo/replication-v2.rst

Your suggestion for changing the semantic of the API is interesting – I would 
interested to know what others in the community thought about this as well. 
Maybe you could file a BP and suggest this improvement?

Regards,
Doug

From: 陈迪豪 <chendi...@unitedstack.com<mailto:chendi...@unitedstack.com>>
Reply-To: OpenStack List 
<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>
Date: Thursday, August 13, 2015 at 4:12 AM
To: OpenStack List 
<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>
Subject: Re: [openstack-dev] [trove]Implement the API to createmasterinstance 
and slave instances with one request

We have read the code of "replica_count" and it's like what I thought.

We have an suggestion to extend this feature. When users set "slave_of_id" and 
"replica_count" at the same time, we just create replica instances. If they use 
"replica_count" without using "slave_of_id", we should create an master 
instance for them and some replica instances of it.

For example, "trove create $name --replica-of $id --replica_count=2" will 
create 2 replica instances. And "trove create $name --replica_count=2" will 
create 1 master instance and 2 replica instances.

What do you think Doug?

Regards,
tobe from UnitedStack

------------------ Original ------------------
From:  "陈迪豪"<chendi...@unitedstack.com<mailto:chendi...@unitedstack.com>>;
Date:  Thu, Aug 13, 2015 12:25 PM
To:  
"openstack-dev@lists."<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>;
Subject:  Re: [openstack-dev] [trove]Implement the API to createmasterinstance 
and slave instances with one request

Thanks Doug.

It's really helpful and we need this feature as well. Can you point out the bp 
or patch of this?

I think we will add "--replica-count" parameter within trove create request. So 
trove-api will create trove instance(aync create nova instance) and then create 
some replica trove instances(aync create nova instances). This is really useful 
for web front-end developers to create master and replica instances in the same 
time(they don't want to send multiple requests by themselves).

Regards,
tobe from UnitedStack

------------------ Original ------------------
From:  "Doug Shelley"<d...@tesora.com<mailto:d...@tesora.com>>;
Date:  Wed, Aug 12, 2015 10:21 PM
To:  
"openstack-dev@lists."<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>;
Subject:  Re: [openstack-dev] [trove]Implement the API to create masterinstance 
and slave instances with one request

As of Kilo, you can add a —replica-count parameter to trove create —replica-of 
to have it spin up multiple mysql slaves simultaneously. This same construct is 
in the python/REST API as well. I realize that you still need to create a 
master first, but thought I would point this out as it might be helpful to you.

Regards,
Doug


From: 陈迪豪 <chendi...@unitedstack.com<mailto:chendi...@unitedstack.com>>
Reply-To: OpenStack List 
<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>
Date: Tuesday, August 11, 2015 at 11:45 PM
To: OpenStack List 
<openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>>
Subject: [openstack-dev] [trove]Implement the API to create master instance and 
slave instances with one request

Now we can create mysql master instance and slave instance one by one.

It would be much better to allow user to create one master instance and 
multiple slave instances with one request.

Any suggestion about this, the API design or the implementation?
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to