And here are a few comments on the design. I'm not familiar with the
odyssey design so am not able to comment on some of the details involved.
--------------------------------------------------------------------------------
FEEDBACK ON TECHNICAL DOCUMENTATION/CODE
--------------------------------------------------------------------------------
Reviewer Name : Nick Solter
Document/Module Title : Design Specification Script Based Plug-in
Controlled Data Replication in Odyssey
Document/Module Author : Tim Read
Document/Module Version/Date : 0.4 7 August, 2008
Reviewer Preparation Time :
------- ----------------------------------------------------------------
No. Location Sev. Comment
------ -------- ---- ----------------------------------------------
NAS-1 Page 10 medium
Maybe you'll get to this later, but I'll ask anyway. It looks like the
SCB specification is properties of the PG. Sort of like GDS. Is there
any facility for writing a permanent SBP configuration, sort of like an
RTR file for an agent?
NAS-2 Page 10 medium
The choice to make each program need to be able to validate is
interesting. The choice for agents was to provide an explicit VALIDATE
method. Did you consider doing that? Why wouldn't that work here?
NAS-3 Page 12 low
It sounds like the mechanism for the callback methods (programs) to
communicate with the framework is by printing stuff to stdout. How will
this coexist with error/warning messages to stdout? Have you considered
some sort of API similar to the SCHA API for agents in SC?
NAS-4 Page 12 low
Why can't the framework create and destroy the resources/resource groups
instead of making each script do it?
NAS-5 General
I'd like to see a table calling out exactly what someone would need to
supply and commands to run to plug in to the framework. I'm still
slightly unclear on that. A sample application would be great as well.
Thanks,
Nick
