Re: [Networker] storage node control from server

On Tue, 15 Jul 2003 11:30:48 +0200, Frank Altpeter <f.altpeter@BROADNET-
MEDIASCAPE.DE> wrote:

>[networker-6.1.3 on linux-2.4.21]
>
>i have set up a storage node these days to redirect cloning to this
>machine, which is physically located in another data centre.
>
>I'm a bit confused about how to control and check the storage node.
>
>The system is configured on the master server (r2d2) as a storage
>node (c3po) with a ADIC "camping fridge" (Scalar 100 jukebox)
>providing three DLT8000 tape drives.
>
>According to nsr_storage_node(5) the name is rd=c3po:ADIC, but i
>seem to be unable to contact it:
>
>[root@r2d2 root]# nsrjb -j ADIC -C
>nsrjb: error, No jukebox named `ADIC'
>[root@r2d2 root]# nsrjb -j c3po:ADIC -C
>nsrjb: error, No jukebox named `c3po:ADIC'
>[root@r2d2 root]# nsrjb -j rd=c3po:ADIC -C
>nsrjb: error, No jukebox named `rd=c3po:ADIC'
>[root@r2d2 root]#
>
>If i look via the nwadmin gui, i see under [media/jukeboxes] menu
>the entry for 'rd=c3po:ADIC' listed, and i also can work with the
>jukebox and its drives via the GUI.
>
>So, what am i doing wrong? I thought that all tasks can be done from
>the master server, and there is no need to login to the storage node.

Not quite. An nsrjb command needs to run on the storage node that controls
the library, so in general you do need to login to the storage node.

Strangely a standalone device can be controlled in this way, e.g.
nsrmm -f rd=c3po:/dev/rmt/0cbn -m
but nsrjb does not allow this.

There is a way though. NetWorker uses the nsrexec/nsrexecd communication
stream to control storage nodes for its internal operations and when using
nwadmin. You can use this too, using the following commands as root on the
NetWorker server:

export RUSER=root
export RCMD="nsrjb -s r2d2 -C"    # optional -j rd=c3po:ADIC
nsrexec -c c3po

You can of course write a short script to make this easier.

Using this method to run a -C command is technically a little long-winded.
You're asking the server to ask the storage node to query the server for a
volume list then send it back to you on the server. The information is of
course already on the server and available by nsradmin. This only becomes
significant when you use a script to query multiple large jukeboxes on
storage nodes, but in that situation the difference in performance is
significant.

Incidentally "master server" is not a term used with NetWorker, it's a
NetBackup (spit) term.  :-)

--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=