I think several of the developers have seen your issue, a view indexing performance hit of about 25%.
Git bisect indicates this began at commit f0230527 https://github.com/apache/couchdb/commit/f0230527 I was thinking of reverting it (with some gnarly conflicts), and perhaps break it into logical sub-commits to identify the source. Filipe, what do you think? Script to test: http://friendpaste.com/1Z1pBlZ8x4GemalOxBvYQq On Fri, Feb 24, 2012 at 7:32 AM, Stefan Kögl <[email protected]> wrote: > Hi, > > On Thu, Feb 23, 2012 at 1:26 PM, Bob Dionne > <[email protected]> wrote: >> could you possibly try using this release candidate? >> >> http://people.apache.org/~nslater/dist/1.2.0/ > > I've now installed the release candidate of the 2nd round and beam.smp > and couchjs are currently maxing out the CPU. The view is not fully > built, but I'll keep monitoring it and report back if the performance > decreases at a later stage. > > The design document is available at > > https://github.com/gpodder/mygpo/tree/master/mygpo/users/_design > > but the data is not public. > > > -- Stefan -- Iris Couch
