I think there is no general answer to your first question. Both priming strategies have advantages and disadvantages. Random primed cDNA will reflect all RNA molecules (intact and fragmented mRNA, rRNA) and the cDNA will not show a 5´-3´ bias (i.e. under-representation of 5´sequences of long transcripts). Moreover random priming usually results in larger amounts of cDNA and is required for transcripts without poly-A-tail (histons). In poly-dT-primed cDNA you have a higher representation of (intact, processed) mRNA- this may be important if you want to be sure to detect transcripts of living cells. poly-dT-primed synthesis should also be used for the cloning or analysis of complete transcripts.
I do not see a real advantage to mix both priming strategies (except you will not be confused about used primers, if you switch between them) but it will work.
Using a combination of both is generally recommended as using just oligo(dT) can cause 3' end bias. Most commercial SuperMixes will provide both. Good luck!
many kits such as Bio-Rad Iscript reverse transcriptase kit, use mixture of both primers do reverse transcripte whole long RNA. These kits are generaly create cDNA for real-time PCR that requires short amplicons. So you can mix them with reducing the consantrations. But it is not adviced that to use mixture of primers for a cloning experiment or any experiment that requires full lenght RNA. Due to the fact that mixing them will also cause short RNA fragments too.