Well [http://people.planetpostgresql.org/xzilla/index.php?/archives/103-PostgreSQL-fails-a-Red-Hat-user.html that] certainly caused a stir didn’t it? While I do tend to agree with a lot of the points David and Merlin made (Generally I think PostgreSQL will outperform $QL $erver, that Oracle is a real PITA to deal with, and that you can get good support within the PostgreSQL Community but there are no garauntees), in this case I think I need to speak up for Sean in this conversation since some of the criticism toward him is completly off-kilter; the two primary things being that people are overlooking that he is not a developer on the project and that it was Red Hat, not Sean, who suggested running Oracle. Now, I have spoken with Sean in private and he agreed to share the links with me on the condition of non-disclosure, so I can vouch that the posts are real and the problems did exist. At least to a point, remember that much of his story relates to the interaction between the application developers and Red Hat services employees, so in that case we really don’t have any way of getting details on that, but I’ll take Sean’s word on things. And generally I think most of us here would agree with some of his basic points:
PostgreSQL is not a golden key for making things scale, you do need to throw hardware at big problems. (I think it’s less hardware than some other databases, and the money you save in licensing can often be put toward hardware, but still…)
You should not expect the community to do your job for you. People talk about how great open source support is and in my experience it is better than commercial support simply because there are no strings attached. I remember having a support contract for Cold Fusion once that I was not allowed to make use of without the approval from my department head, since we were limited on the number of incidents that we could report in a given year. This seemed ridiculous to me when compared to help I recieved from the PHP community, where I could ask question after question if I wanted. Now, that said, if you’re planning on building a multi-million/billion dollar business, you shouldn’t expect that all your problem will be solved through some free mailing list. If you find bugs in the software then you will be help regardless, but if you need help with setup, configuration, or training, it’s silly to even try to do it without commercial support.
People need to be able to discuss issues without divulging details, and we need to be able to accept those criticisms in context and provide coherent responses to them. Ok, his post wasn’t really about that, but there is a tone of that in his responses, and I think some of us get a little carried away when people start comparing PostgreSQL negativly to other systems.
I still think that PostgreSQL offers these folks the best solution, because even though they’re having troubles, they have the option of switching to another support company, an option you just don’t have when running Oracle or $QL $erver. It’s also worth repeating that there are people running systems that large on PostgreSQL, so from a database software point of view it can surely be made to happen, we just need to find a way to get more of that information out to the community.