Re: [Maposmatic-dev] Database update becoming critical

[David Decotigny]

Hello,

Maxime Petazzoni wrote:
> A
> very simple hourly crontab could be created from this. We'd have to
> check of course that this process has a significant time advantage. I
> believe the more frequent we make the update process, the faster it
> could run: I don't think the processing time grows linearly with the
> update interval (aka it takes way longer than 60 times what it takes to
> process a minute diff to process an hourly diff).

It would be very interesting to have some benchmarks about this. At
first I was dubitative about that claim, but on second thoughts, maybe
it's right (startup overhead higher, but working sets smaller on the
other hand => scan/search/sort overheads perhaps smaller). The only
closely-related benchmark I'm aware of is
http://lists.openstreetmap.org/pipermail/talk-fr/2010-May/021293.html
(thanks, Thomas), do you know of any other tests with osmosis in the loop ?

To test this... Is there a "COW" feature for pgsql which would allow us
to run a few tests with the prod DB, without altering the data delivered
to ocitysmap ? If not, I'm afraid we'd have to test on another machine.

Regards,