Thanks Mark. I realize it’s a huge migration. I will be exhaustively testing everything. I like the idea of creating an empty database and pumping the data over. Is there a tool for this? (Actually just found one from Clever Components)
I have looked at the migration docs for the 2.x versions and will do more. Do you think it’s worth me going to 2.5 first or go directly to 3.0? Both will be a ton of work so I’m thinking directly to 3.0. Thanks again! Kevin > On Feb 13, 2019, at 9:05 AM, Mark Rotteveel m...@lawinegevaar.nl > [firebird-support] <firebird-support@yahoogroups.com> wrote: > > On 13-2-2019 16:58, Kevin Stanton kevin.stan...@rdb-solutions.com > <mailto:kevin.stan...@rdb-solutions.com> > [firebird-support] wrote: > > I’m starting to look into migrating from 1.56 to either 2.5 or directly to > > 3.0. I’ve bought the Migration Guide but haven’t read it yet. > > > > It concerns me that I would/will have to recompile all SPs and triggers.. > > Is there an automated way to do this? I have 909 SPs and 785 triggers. This > > will be one heck of a task with my installation base. > > The choice is to either carefully and exhaustively verify if all your > stored procedures and triggers continue to work correctly, and only > recreate those that you know need it, or recreate the whole shebang (you > will still need to carefully and exhaustively verify them). > > Firebird itself has no way to recreate stored procedure and triggers > automatically, although it is possible to craft scripts that rebuild it > based on the database metadata (assuming the source blobs haven't been > cleared). I would recommend against doing that automatically and instead > you should use known good and verified scripts for the actual upgrade, > if only to avoid running into issue with things that are now no longer > syntactically valid or unquoted use of identifiers that are now reserved > words. > > Some people advocate that you shouldn't upgrade a database at all, but > instead recreate it and pump the data over. In a situation with a lot of > deployments that might also fix problems with diverging structure (but > if that is the case, the migration path will be more prone to errors). > > Especially when taking the big step from 1.5, it might make sense to > take it as an opportunity to improve your database design, and get rid > of things that are no longer needed (eg use built-in functions instead > of UDFs, switch from dialect 1 to dialect 3 if you are still on dialect > 1), etc. and then migrate the data from the old to the new. > > Firebird 1.5 to 2.0 was already a big step with a lot of new features, > but also some breaking changes, sometimes subtle, sometimes not. Don't > expect going from 1.5 to be simply backing up and restoring, especially > not with 909 stored procedures and 785 triggers, and more so if you use > a lot of UDFs, or relied on some of the ambiguous behavior that is no > longer valid, used unquoted identifiers that are now reserved words, etc. > > And realize that Firebird 2.0 was released 12 years ago. A lot of > migration knowledge from 1.5 to 2.0 may have been forgotten since (or at > best not easily recalled). You should really carefully study the 2.0, > 2.1, 2.5 and 3.0 release notes and the Firebird 2 installation and > migration guide, and as you bought it, Carlos' Migration Guide to > Firebird 3. > > Mark > -- > Mark Rotteveel > > [Non-text portions of this message have been removed]
