From: Alex T. <ale...@in...> - 2010-07-07 16:32:40
|
Thank you Justin, this is very helpful. On 07/07/2010 11:59 AM, Justin Deoliveira wrote: > It really depends. I would say no there is no inherent assumption in > geotools itself about how a backend will handle multiple concurrent > connections. > > However practically anyone that intends to use GeoServer in a server app > like GeoServer (or a multi threaded client app for that matter) is > essentially making the assumption. And the geotools datastore api does > imply that implementations support multi threading. > > I have this same issue for H2. I intend to enable the MVCC by default soon. > > -Justin > > On 10-07-07 8:53 AM, Alex Trofast wrote: >> Greetings all, >> >> I had a quick question regarding GeoTools and multiple transactions. >> When running some of the transactional concurrency tests I would get >> failures and/or hangs with Ingres. However when I turned on MVCC >> features these tests would pass/not hang. Does GeoTools ASSUME that the >> DBMS is MVCC enabled? This is actually a very new feature in Ingres and >> is still in beta so I'm just wondering if there is any control to what >> the expected behaviour is. >> >> Thank you kindly, >> > -- Alex http://bit.ly/geodb |