Re: Oracle 10g (ASM vs Veritas Clustered FS)

> You don't necessarily have to go running to the vendor
>everytime something goes wrong. Infact, it's better if you don't
>have to. Vendor response times can vary and you really don't
>want to be depending on that.

Depends on the "you." For a gross overgeneralization, what I've seen in places that decide to do things themselves to save on recurring support fees, everything goes fine for a year or three, then it all goes to hell as the cumulative lack of preventive maintenance finally tsunamis. Like no one remembering to check when the batteries in the UPS were installed. Or employee turnover. Or suddenly discovering that O7 don't work no more.

> This should not concern the Oracle portion of the system much.

>Oracle is actually pretty good about tolerating people being too cheap

>with hardware. The data should survive even if no work is getting
done.

In general, I agree with this. But I have seen specific instances of it just getting f*&^&*(% by repetitive bounces or worse, hardware corrupting data. That's when there _should_ be finger pointing. And I think we might see more problems as Oracle attempts the impossible, increasingly complicated systems on increasingly crappy hardware. The computer equivalent of those stupid "radio controlled" toys from China that go backwards until you press the button and it goes forward in a circle.

jg

-- 
@home.com is bogus.
http://www.pacifict.com/Story/