Subject Re: [firebird-support] Re: Deadlock during index activation
Author Alec Swan

We have thousands of client applications using embedded Firebird 2.1.2 and
one of the reported an index corruption problem which I describe in this

The easiest way to fix corrupted index is to ACTIVATE it and that's what the
maintenance plan is doing. We could make the maintenance plan smarter and
only activate index if gfix reports it as broken, but that's just an
optimization. The bottom like is that we need to activate indexes to fix
index corruption.

However, it's good to hear that you have success with your large deployment.
What version of Firebird are you using? Do you do any kind of maintenance on
your databases?



On Thu, Jul 29, 2010 at 9:42 AM, Leyne, Sean <Sean@...>wrote:

> Alec,
> > Thank you for the explanation. Our application has a maintenance plan
> which
> > calls ACTIVATE INDEX on every index in the database. It is possible that
> other
> > queries and statements are run against the database during the
> maintenance
> > plan. According to your explanation those concurrent modifications may
> > cause the "deadlock" error to be thrown during index activation. Would
> you
> > suggest preventing concurrent queries and modifications during index
> > activation?
> >
> > Where can I get the full message for the "deadlock" error I am seeing?
> (What
> > I reported was the entire message I receive from Jaybird.)
> The have 100+ servers deployed at client locations (with databases between
> 1GB and 19GB in size) and we *never* need to perform index maintenance.
> I would suggest that the "maintenance" plan is over-thought.
> Sean

[Non-text portions of this message have been removed]