Alok,
Nice document. Just a few comments.
Jean
Section 2.0
You mention that all checkpoints will store and retrieve data via the
DOC. Per the discussion this morning you should also state they should
provide the interface to
not use the checkpoint interface.
Section 2.2
I believe TD will be used to discover the existence or not of the dc
dataset.
TI/TD/Transfer: You also depend upon them providing the code to create
their entries in the DOC. i.e. the TransferInfo, TIInfo, TDInfo classes
which have to to_xml, from_xml and other code the DC will use to
actually create the nodes.
section 3.0 diagram
Here you actually show target discovery, however if the checkpoints are
meant to be visually in order I think that needs to move up to before TI.
Where are the copious amounts of checkpoints that used to be finalizer
scripts represented in this diagram?
What are the post installation modules and data collection modules?
section 3.5
If we're making what used to be finalizer scripts into checkpoints, I
think you should call this the checkpoint section of the manifest.
section 3.6
There's nothing here. I suspect you're still thinking about it?
section 3.7
Be careful with picking running the system binaries. The reason that
some of the binaries were from the build area was that it broke
otherwise. I think, the pick should be to use the binaries from the
build area but you will need to look into this more.
Running the same version of Solaris may not be good enough, we were
running S11 and building S11 and still had issues.
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
