Anyone know why Teradata CS recommends applying FALLBACK to all databases (and hence tables) on vitual servers they support?
It seems an overhead on a small development server that couyld be avoided.
In TDVM environment, the FALLBACK feature is enabled by default and can't be override. This is to recover the lost data from an AMP. There could be cases, where pdisk might get lost due some hardware failure, which might occur on customer's VMWare environment. In such cases, having FALLBACK option, we have chances of recovering data for AMP from its clustered FALLBACK AMP.
Ignoring for now the pro's and cons of using FALLBACK, you said " the FALLBACK feature is enabled by default and can't be override."
I thought this could over-ridden using DBSControl?
Assuming that you have to have FALLBACK on, then you could:
- use BLC to save disk space
- try using BLC against only the FALLBACK copy of the data (look in DBSControl / Compression group) to balance the saving in disk space against extra CPU for decompression processing.