The long processing time was likely caused by large amounts of transparency and graphics used which needed to be processed by the RIP.
If the job failed to print, you may have conflicting settings; "Fold" combined with "Main" Output Tray (although the print features window would usually alert to this conflict).
The size of the RIP'd postscript probably became to large for the system memory.
Try to duplex pages 1&2 followed by pages 3&4 followed by pages 5&6. That would eliminate the question of a possible bad pdf source file.
You could flatten the transparency in the pdf, that would speed up the spooling (save a Press or High Quality pdf as opposed to x1a)