On 15 December 2011 12:32, sly (sylvain letuffe) <[email protected]> wrote:
> Will the maintenance and stress on this server (therefore on it's sys admins) > whould be painless if all read api calls were directed to other servers ? Let me reply to this in slightly hand-wavy terms. Most of the load on our core infrastructure comes from read requests. Anyone who can use another service, which gets its updates via the minutely replication service, helps make the core go faster. This counts especially for all those things that make requests against the main API for non-editing purposes. Also, we (the Operations Working Group) are aiming to spread core postgres read requests over multiple replicated servers. We're working hard on getting the relevant hardware in place to start trialling this out, but it's a big project. The original topic was about replication for rendering, so a comment on that - that's not something we're tackling centrally. We have the minutely diffs to allow robust, low resource replication out of the core infrastructure, and we don't have plans to change this. As Frederik has already said, if you want to replicate rendering databases, jxapi databases or anything else, set one machine up to read the diffs, and replicate from that using whatever system works. It can be done today, if anyone wants to do so. Cheers, Andy _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
