Hi Jane,
I will say what almost everyone else will say - It depends on a lot of
things....
I can give you some indication of what I did last weekend, although it was not
a unix
client the principle stays the same.
My setup was as follows:
ADSM server: OS/2
Client: Novell 3.12
10530 files / 700 Meg
Network: 4M/bit Token ring
It took me 1 hour to back up the complete Novell machine with all the client
data
stored on dasd on the OS/2 server. (This will make a big difference when you
do your restore). If you look back in the posts you will see a lot of posts
regarding restores taking a very long time etc.
My restore took just slightly longer than 1 hour.
From the above you can see that the backup and restore took about the same
time - but again there were no tapes involved.
As far as your question about the file systems are concerned:
What kind of network do you have in place?
If you have a reasonable network then you can do your restore on a file system
basis - by that I mean running a restore command for each
filesystem(directory).
This is commonly referred to as running more than one restore stream - which
will speed up your restore quite a bit.
As far as I can see you will be able to do your restore with not too much
trouble.
Will you have a working system as far as the Operating system and the network
is concerned? If so, it is just a matter of loading the adsm code and setting
the
options file with the correct information in it.
Hope this helped,
Regards
Christo Heuer
Johannesburg
South Africa
Christoh AT absa.co DOT za
> Next week we are planning on restoring a node (server). This is not the
> server that has ADSM on it which is ADSM MVS mainframe.
>
> The situtation here is that we have a Sun Solaris server in a test
> environment containing People soft data which I will call A and a test
> server, same type and disk space ect. that is empty which I will call B
> and will use B to restore an inactive versions on the data stored backed
> up from A.
> Our Operating Support support plan on shutting down A and using the IP
> address that A was assigned. They want to restore the OS once they
> are done doing their thing, I will aplly the ADSM client piece on the B
> machine and restore the data.
> There are appox. 40,000 files, and takes about an hour and a half with
> some logging going on during this time to complete the backup of A. I am
> not sure how much Meg is being transferred offhand, but from other
> poeples experiences, how long would it take to restore data that took
> say 50 minutes (without logging) to be restored?
>
> When I display the node name of A there are 11 file space names listed,
> they are: /, /usr,/var,/export/home,/opt,/u01,/u02,/u03,/u04,/u05,/u06.
> Now does the file space names mean that they are the actual directory?
> By looking at the manual ''User Guide and Reference for Unix", ( I think
> this manual has be updated and may be called something else now) I
> plan on restoring the server by one of the following ways:
> by file specifications
> directory tree
> subdirectory path
>
> Now is their a better way to restore this machine? How many and
> which other items I mentioned have a logic error in it which I am not
> aware of indivually but I am assuming that in all likely hood that there is.
> Thank you for any hints, insights
>
> Jane Dagostino-Snyder
> University of MN
|
