launchpad-dev team mailing list archive

[Robert Collins <robert.collins@xxxxxxxxxxxxx>]

On Wed, Mar 30, 2011 at 5:08 AM, Francis J. Lacoste
<francis.lacoste@xxxxxxxxxxxxx> wrote:
>
> On March 29, 2011, Tom Haddon wrote:
>> > What-do-you-all-think?
>>
>> I think my primary interest here is in what the practical implications
>> of this will be in terms of the number of "downtime/DB schema upgrade"
>> deployments there will be. Is what you're proposing going to mean more,
>> or less rollouts, and what's the predictability of those going to be?
>>
>
> For the time being, there is no change in that respect. We are still going for
> one DB deployment a month based on the existing calendar.
>
> We are going to work on removing the friction around the deployments and
> toward reducing the overhead of doing them. Once we have very little friction,
> (basically, DB time + minimal overhead), we'll want to target to doing them
> once a week. But we are not there yet. So short terms, nothing really change
> except for all the other details in Robert's email which boil to:
>
> * Stop using milestone
> * Cancel the ReleaseManagerRotation process.

Right, I was trying to include the big picture as well, guess I let it
get a little fuzzy :(.

In the future, once we've got downtime deployments down to a fine art
(https://dev.launchpad.net/LEP/ReliableDBDeploys) we can look at
signficantly more downtime windows, but much shorter and more
controlled ones. That has a bunch of technical challenges involved in
it - for now we can remove a bunch of overhead that formed around the
concept of once a month, feature-including, high-risk deploys.

-Rob