On Thu, Feb 23, 2012 at 1:11 PM, Peter van Hardenberg <p...@heroku.com>wrote:
> My hunch is still that your issue is lock contention. > > How would I check that? I tried looking at pg_locks but I don't know what to look for. > We have many customers who do much more than this throughput, though > I'm not sure what level of resourcing you're current at. You might > consider experimenting with a larger system if you're having > performance problems. > > Heh. I thought you might say that. :) It's definitely worth considering, but as youmight expect, I want to exhaust other options first. For customers who do much more (or even comparable) throughput, can you tell me how big of a system they require? Also, as per Andy's suggestion, I'd like to try doubling checkpoint_segments. However, it appears that that is one of those variables that I cannot change from pgAdmin. I don't suppose there's any way to change this without rebooting the database?