That's pretty much what I figured and others have likewise indicated
that this is the case. So, as it has been suggested, it will be
necessary to configure the current storage node so it can see all the
hosts before making it the primary server. But, this raises a few
questions that I'm unclear on:

1. When you say "metadata", I assume you mean the information on the
files for the index itself. As such, isn't it true that that the "real"
data is sent to the storage node to be backed up by the devices attached
to the storage node, and the server never actually sees this "real"
data.

2. At what point does the server receive the "metadata"? Is it
throughout the backup or once the client has actually finished writing
its "real" data to tape.

3. As near as I can tell, the 'nsrmmd' daemon (one per device) and the
'nsrexecd' daemons are the only daemons that run on the storage node, so
I guess these handle the backing up of the data from the client to the
storage node. I've only seen the 'nsrindexd' daemon running on the
server. I guess this handles the updating of the metadata for the
indexes? Would make sense that this cannot be done on the storage node
since that deamon never runs there, correct?

Thanks for your help.

George

Carl Farnsworth wrote:
>
> This won't work - the client must be able to communicate directly with the
> server.  Not only is the index updated by metadata communicated directly
> from client to server, but the actual save is instigated by the server
> sending the save command to the client.  Also, the client has to be
> validated against the server directly.
>
> On Tue, 8 Apr 2003 09:42:33 -0400, George Sinclair
> <George.Sinclair AT NOAA DOT GOV> wrote:
>
> >Hi,
> >
> >Is it possible for the primary server to update a client index if it
> >cannot communicate with the client but the storage node can? In other
> >words, if the primary server can communicate with the storage node and
> >the storage node can back up the client, can the primary server update
> >the client's index via the storage node? Clearly, it could write the
> >index to tape since the index resides on the main server, but I'm not
> >sure how the index would get updated while backups are running.
> >
>
> --
> Note: To sign off this list, send a "signoff networker" command via email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=