About the only solution I can suggest is to avoid working for such companies, but I'm be preaching to the choir to say that.
Because a properly installed and maintained Oracle is pretty reliable and straightforward to use. If you even can run your application on Postgres then you wasted your money buying Oracle; you aren't using it for what it's for[1].
[1] I am aware that convincing people that they might need to in future is key to Oracle's business model, but that's beside the point.
Oracle (the database) is meant to be the integration point for Oracle (everything else). If you aren't running any of Oracle's other products, then it boils down to a pure technical decision - do you need feature X? Is it cheaper to buy it or build it? For example do you use OpenFiler or buy VVM or use ASM that comes with Oracle since you need a database anyway? Or maybe you have exotic spatial data needs. Or you want virtual private database/label security (http://wiki.postgresql.org/wiki/RLS - Oracle has had this for years, the Postgres guys are still talking about it, the MySQL camp don't even know that there is such a thing).
Now, I am a long-time Oracle user, and even I say, unless you know and can say specifically why you need Oracle, then you should use Postgres or SAP/DB (or in place of Oracle XE, maybe even SQLite). But if you do need it, then it's Oracle or DB/2, choose your poison.
I guess you're stuck with it if you need to run Oracle financials or Peoplesoft or something like that but one of my professional goals is to never work for another company big or dumb enough to use either.
Oracle (the database) in the hands of anyone basically competent is easily capable of 5-9s reliability. Not that I'm saying Postgres isn't mind. I'm just saying, use the right tool for the job (and hire the right people).
I find DataPump very easy to use, and very fast, it's I/O bound even on serious storage arrays. YMMV. SQL*Plus is definitely showing its age I agree.
Having said that, I don't think Oracle was totally neglecting that area. They had a few decent free tools available (SQLDeveloper, JDeveloper, trial DB versions) and while their articles didn't come close to the sheer amount IBM offered, there was enough information out there.
They could've marketed that better, though. But in that regard, they're certainly not much worse than a lot of other enterprise companies. You can't measure everything by the way Microsoft and Apple are doing it.
Might as well make the best of it.
Not that this is the case everywhere and everytime, but if technical decisions were left solely in the hands of people that lack technical skills, the situation would be even worse that it is now, with all the "enterprise software" floating around and causing trouble directly proportional to their cost.
Every few days on pgsql.general mailing lists, there is someone asking for a good Postgresql book (the "High Performance MySQL" kind)
Less books = one more tick mark for Oracle.
If you won't do it, why should anyone else?
My take is that the overhead of maintaining and using Oracle means that you have to cut resources somewhere else. I'm pretty sure a competent DBA and developers can ensure atomicity and integrity using Postgres (instead of Oracle), and have a lot of time left over to create more features in the application.
Its just that RAC is a special technology with no equivalent on other DBs. It is usually placed where you need load balancing on the DB level. It normally has very high availability that is a good fit for mission critical DBs. I hate seeing good technology wasted, but you are right that if Postgres is already there and works well, there is no point in moving anywhere.