Sarah Jelinek wrote: > Hi Sanjay, > > A couple of other comments/question on the requirements document: > > Replication Goals: >> System software on the replicated system must be identical to the >> master system while >> accounting for minor differences in hardware and configuration >> parameters. This applies to bare >> metal and virtual machines. > > Today with flash the default is to get all of the software on a system > in the archive: system software, application software... Customers use > this to create a golden master to deploy systems in their data centers > after having tested and certified the applications on a version of > Solaris. They can choose to exclude data. In reading this goal it > isn't clear to me what you mean by system software. Based on this goal > and requirement #4 I assume system software is: > > software installed on current root dataset, i.e. > /rpool/RPOOL/be1 > and > zones datasets? Yes that is correct. I will clarify this. > > Assumptions: >> 2. No user data payload. Flash provided the ability to include user >> file systems. This was >> reasonable when disks were much smaller and data was closely tied >> to the system. Today, most >> user data is either on a NAS or SAN device and they can be huge. >> User data is often separately >> backed up. Hence it does not it does not make sense to include >> user payload. > > There could be user data in the root dataset so we cannot say we will > exclude user data. We have no way, unless we interpret the data under > the root dataset to know if it is user data or not. In general, user > data won't be in the root dataset, but it could be. > > What constitutes user data payload? Does this include application > software and configuration. Or, anything that isn't in the root > dataset, or root and zones datasets? Or datasets not in the root pool? > I am just trying to clarify this. I was specifically contrasting Flash functionality, where the user has the ability to add other file systems. In this context it could mean other BEs in either the root pool or other pools
> > I think we should have some specifics about what we intend to support > and what we don't intend to support in the requirements or goals. > > If the answer is we think user data is anything not in the root > dataset and zones datasets then it seems to me we have to interpret > the data in the zone configuration file to know what datasets to copy. > Unless we copy everything. We have to try to figure out what datasets > are used for what if we exclude any data in the root pool. > > sarah > >
