I'm trying to get my brain around parallel operations, and how
the various settings and options control parallel operations.
First, I guess it would be helpful to define my terms.
I consider there are four different things can
be done in parallel during a mirror operation:
A) examine timestamps of files within multiple directories
over multiple control connections
B) transfer X files in parallel over X data connections
C) transfer segments of a single file over X data connections
D) combination of (B) and (C), in which segments of
X different files are transferred over Y data connections,
and Y > X
I am unclear on how the different settings and switches
relate to the possible parallel data operations.
For example, I can see that setting
mirror:parallel-directories on
mirror:parallel-transfer-count ${parcount}
or using
--parallel=${parcount}
both have the effect of implementing (A).
But it is not clear to me how these parallel settings
affect file transfer during a mirror command.
Can you please clarify if and how parallel file
transfer is used during a mirror operation
If someone could help me understand how this all
works together I would be grateful.
Thanks and Best Regards,
Jeff Buckles
Aloha, OR, USA
