I'm reviewing ingres to see how it can be placed in a LA?P environment. The feature list looks great and the installation was easy.
Now next to the tests i'm going to be running myself i would like to hear some opinions from people who have been working with it. Since our little community is pretty well known for its assertive and critical nature, I was hoping some here might have some experience with it and could share them.
A shameless Ingres bigot speaks
As a (waaaaaay) long-time user of Ingres and a keen follower of TDWTF (I am drinking coffee from my WTF mug as I write), I really appreciate Ingres' willingness to do things right even if it meant resisting demands for mis-features. This did Ingres no commercial favours, but by golly, if you worked with it instead of against it, you would end up with trouble-free, self-managing systems.
In recent years the attitude has relaxed a lot, and some of the things that were previously shunned as being wrong-headed have been adopted. I am past caring. I don't have to use them, and I make a decent living picking up the pieces when others are seduced onto the rocks.
In terms of reliability, recoverability, upgradabilty, and supportability Ingres is a bet-the-business open source product. It is highly standards-compliant both in terms of syntax and semantics. There are a few silly things with it, like it ships with a default isolation level of serializable when everyone else defaults to read committed, but they are easily adjusted.
I happen to have access to some very highly placed sources in Ingres Corporation, and I hear that there are going to be some "eye-popping" announcements at the UK Ingres Users Association conference in London next month. Take a look at http://www.iua.org.uk/docs/IUA_Invitation_2009.pdf for details.