| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 387.1 | | TOOK::STRUTT | Colin Strutt | Mon Oct 08 1990 16:45 | 13 |
| I;m not sure I fully understand your question, nor exactly why you are
asking it.
However, there is some information in the SRM, particularly chapter 2,
which describes the MIR in terms of different categories of information
that need to be maintained. Each of these (four) categories has its own
storage, distribution and access characteristics - hence each can be
treated separately.
Perhaps you would like to ask a more specific question if the above and
the SRM are not sufficient?
Regards
Colin
|
| 387.2 | Non-Interrupted Mgmt | SUOSW4::HARDT | Daniel Hardt,EIS/Germany,765-5286 | Tue Oct 16 1990 12:46 | 20 |
| RE.-1:
The SRM does not answer my question regarding data-mgmt behaviour
within DECmcc.
I'd like to be more specific:
Assume you have a nw-mgmt scenario with a two-system NCC (resp. DECmcc) in
place, one of the systems also considered to be the backup of the other.
How can one guarantee permanent access to the NCC's entire
data/function-base without having manually doubled all the relevant
data ?
The SRM is dealing also with things like RMS and Global Sections, but
this is neither system-independent (MIR specific) nor safe per-se.
I would think of a DECmcc mgmt feature, which allows to define a DECmcc
system operating also as an replicated data-store to another one -
sensefull ?
Rgds.
|
| 387.3 | Use VAXclusters and disk shadows | NSSG::R_SPENCE | Nets don't fail me now... | Tue Oct 16 1990 15:57 | 7 |
| One way is to use a VAXcluster with shadow disks.
Another is to use DFS to hold the specific info and store in on a disk
in a VAXcluster and shadow the disk.
s/rob
|
| 387.4 | | TOOK::STRUTT | Colin Strutt | Wed Oct 17 1990 02:29 | 19 |
| re: .2 and .3
We chose in DECmcc to *not* reinvent the wheel with regards to data
management. Thus, although we can see the need to implement
replicated and/or distributed database capabilities for management, we
would prefer to utilise corporate products to perform these functions.
In .3, Rob points out some ways in which you can separate the data
from the system(s) running DECmcc - that may be sufficient for your
needs right now.
Over time, we would like to be able to provide enhanced data storage
support - it's "on the list" of things to think about in the future.
Regards
Colin
P.S. If you read the latest SRM (v1.1 - coming to a store near you
soon) you will find references to particular operating systems have
been removed - so no more RMS stuff, ok.
|
| 387.5 | Workaround ? | SUOSW4::HARDT | Daniel Hardt,EIS/Germany,765-5286 | Wed Oct 17 1990 12:40 | 17 |
| RE.-3/4:
VAXcluster is a LOCAL solution, NCC's are typically not that way.
DFS does not eliminate single-point-of-failure on the data side.
I don't see a corporate product to eliminate this problem now -
with the exception of a 3rd system being a local dual-host backup
and having DFS'ed files for the remote DECmcc (cost, concurrent
access ?!).
Is it generally possible to copy the according files to another DECmcc's
data base ?
Rgds.
|