Fabian, I would not rush into FB 3.0 on a production system. We did and we regret it. We had the same hopes that the SMP would make great improvements. At this point things are actually running slower for us with FB 3.0. Some of the queries needed to be rewritten to perform better, and we have worked around a FB 3 bug where passing a null in a query param could make it do a full table scan instead of using an index. [#CORE-5196] Index not used when search criteria is NULL in FB 3.0 RC 2 - Firebird RDBMS Issue Tracker http://tracker.firebirdsql.org/browse/CORE-5196
In general FB 3 seems much slower than FB 2.5 with anything that does a full table scan. We have struggled for the past week to try to get performance back to where we were under 2.5. Also not sure if you can revert to 2.5 once you move to 3.0. The GBAK file under 3.0 will not restore under 2.5. Best of luck. --Pete ---In firebird-support@yahoogroups.com, <fabianch@...> wrote : Hi All We have a multi-user, multi-company, cloud based system running with Firebird 2.54 as the main database, with about 100 concurrent connections. We are considering migrating to FB 3.0 to take advantaje of the SMP architecture, we have a 32 core / processors server, so we are hopping to see an increased performance. We also have the need to set up online replication. Would anyone be able to advise if it it too soon to jump into FB 3.0 for production? And if the performance increase would be enough to make the upgrade worth considering 3.0 has not been around for long yet, and may be "taking too much risk" to move into it so soon? And regarding the replication, any advise on what product may be the right choice? Cheers Fabian
