launchpad-dev team mailing list archive

[Gary Poster <gary.poster@xxxxxxxxxxxxx>]

On Jan 11, 2010, at 4:00 PM, Francis J. Lacoste wrote:

> On January 11, 2010, Guilherme Salgado wrote:
>> On Mon, 2010-01-11 at 15:32 -0500, Gary Poster wrote:
>>> On Jan 11, 2010, at 2:59 PM, Guilherme Salgado wrote:
>>>> On Fri, 2010-01-08 at 09:27 -0200, Guilherme Salgado wrote:
>>>>> On Thu, 2010-01-07 at 15:43 -0500, Francis J. Lacoste wrote:
>>>>>> On January 7, 2010, Guilherme Salgado wrote:
>>>>>>> As I explained earlier, to make the read-only mode switching
>>>>>>> transparent, I turned the main_master/slave config variables into
>>>>>>> properties of DatabaseConfig, which return either rw_main_* or
>>>>>>> ro_main_*, depending on the mode we're on.
>>>> 
>>>> When I came up with the plan above I think I blindly assumed the next
>>>> time a request kicked in after a mode switch, it would re-connect to
>>>> the database, thus connecting to the appropriate one. That doesn't
>>>> happen; the connections to the database will be reused (and thus stay
>>>> connected to the DB that was appropriate when they were created) until
>>>> they are explicitly closed or left marked as dead.
>>>> 
>>>> I can't think of an easy way to close a connection at the beginning of
>>>> a request (as we don't want to switch modes mid-request) and only once
>>>> for each connection. Do you guys have any ideas?
>>> 
>>> Can we check to explicitly close connections at the end of a request?
>> 
>> (Just discussed this with Gary on IRC and...)
>> Ok, so the plan now is to, at the end of a request (or maybe at the
>> beginning?), compare the databases of all our connections against
>> dbconfig.main_master and if they differ we close them, forcing a reopen
>> which will then connect to the correct db.
>> 
> 
> Sounds good. I'd do it at the start. That seems also related to the problem 
> with the DisconnectionError we are seeing because of the stuck storm state.

A comment I just added to https://bugs.edge.launchpad.net/launchpad-foundations/+bug/504291 is related, and might suggest a reasonable place to do what we need for the read-only switch.

Gary