Hi,
1a) I have been informed that the disk (DPM storage LUN) IO payload size (disk data transfer block size) (i.e. whilst reading/writing the VSS snapshots) is 64KB for both reads and writes.
1b) Or have I been mis-informed and what my source actually meant... is that the 'disk volume blocking factor' of the storage LUN for DPM is 64 KB, i.e. that the many GPT partitions which comprise used DPM storage space have a blocking factor of 64KB ?
1c) Or, are both statements above true?
2) If (1a) above is true, then is there a way to increase the IO payload (disk data transfer) size, and thus move more backup data per IO ?
3) I've also been told that a DPM server instance might also typically require a large amount of page file space. Can someone explain why this might be the case? Could this be overcome by providing more RAM? Or is there something inherent in the design of DPM as an application that simply requires 'N' x RAM for page file space?
4) The requirement of a large amount of page file space would seem to indicate to me that the page file will experience lots of activity, and thus that the page file for a DPM server instance (physical or virtual) should itself be placed upon the fastest possible storage. True?
Thanks.