[Bucardo-general] master/master with bulk updates, load balanced

Greg Sabino Mullane greg at endpoint.com
Mon Aug 24 20:00:14 UTC 2015


On Mon, Aug 24, 2015 at 11:55:16AM -0700, Florin Andrei wrote:
> I'm running PG in the cloud, currently single instance. I want to
> have more than one instance, to improve reliability and
> availability. I'd prefer a master/master scheme, to increase
> availability of not only the read endpoints, but also the write
> endpoints.
> 
> We do bulk uploads almost exclusively - a Python script waking up
> once in a while, that dumps more data into the DB. Internal users
> may query the DB for data analytics whenever they like.
> 
> Is Bucardo a good fit for this scenario?

It should be. If all the dumps go to a single server, all the better.

> What would be some things I would need to pay special attention to
> while installing / configuring / running the cluster?

All your tables will need to have primary keys or unique indexes. 
If there is a chance for conflicts, you should think about the 
best way to handle them. You might want to use the offset-sequence 
trick [1] as well to help prevent conflicts and for better auditing.

[1] Server A does 1,3,5,7,... Server B does 2,4,6,8, etc.

-- 
Greg Sabino Mullane greg at endpoint.com
End Point Corporation
PGP Key: 0x14964AC8
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 163 bytes
Desc: Digital signature
URL: <https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20150824/780d8e5c/attachment.sig>


More information about the Bucardo-general mailing list