Firebird News

Tuesday, August 07, 2007

Firebird clustering - the bad news

The bad news is that it involves coding in firebird core (at least in the lock manager code)

I'm afraid it won't work. The Classic processes should be able to
synchronize the logical I/O operations (page reads/writes) in order to
perform properly. It's done by the integrated lock manager, but it
cannot work between cluster nodes. As soon as the lock manager is
enhanced (or better say fixed) to delegate the calls to the external
distributed lock manager (e.g. RedHat DLM), this would become a workable
solution.

http://tech.groups.yahoo.com/group/firebird-support/message/88130

.:():..:():..:():.

Firebird SQL Clustering questions

Here comes the fire cluster question
I want to build an Load Balanced Active/Passive Firebird SQL Cluster
(two Active/Passive clusters which have access to the same data). So
the set up will be as follows:
1. Two Redhat GFS Clusters
2. One shared storage of 1TB
3. Layer7 Switch

So my idea is to implement GFS on the shared storage (/shared/db) and
then create two quorums one for one cluster, and the other for the
second cluster. After that I would register the following into a
cluster service:
- Shared IP address
- Shared Storage
- Firebird init.d script

So have anyone managed to create such scenario or is this kind of
scenario supported on Firebird 2.0.1 Classic.


.:():..:():..:():.