Remotely resetting an autochanger 2

[does]

Hi,

Is there a way to remotely reset an autochanger from the commandline?

nsrjb -s backupserver -j rd=storagenode:jukebox -HE

does not work from the backupserver.

thank you

 

[605]

I am afraid that this is not possible. Please read the info i extracted from the nsrjb manpage - i marked the key in brackets &quot;>>> ... <<<&quot; :

DESCRIPTION
The nsrjb program manages resources in two broad classes of jukeboxes, remotely managed jukeboxes and locally managed jukeboxes. Remotely managed jukeboxes are controlled through an external agent. nsrjb communicates with this agent to gain access to jukebox resources. The agent allows multiple applications,including multiple NetWorker servers, to share resources in the jukebox. Examples of agents are
SmartMedia and StorageTek’s ACSLS . >>> nsrjb communicates directly with a locally managed jukebox, there is no intervening agent. <<< Resources in a locally managed jukebox can be used by only one NetWorker server.

For a locally managed jukebox, the jukebox resource is used to track the state of the entire jukebox. The resource records the number of drives and slots in the jukebox. It is also used to track whether devices are loaded, whether there is media residing the slots, and the name of any volume on the media, as well as other information. See nsr_jukebox(5).

 

[DavinaTreiber]

Yes you can do this, but not directly in the way you describe. nsrjb does not accept jukebox and device names in the format &quot;rd=&quot; although nsrmm does (for standalone devices). nsrjb needs to run on the storage node, but you can remotely control this in the same way that NetWorker does. NetWorker sends commands to the storage node via its own communication path, that is from the NetWorker daemons on the server, via the nsrexec program, to the nsrexecd daemon running on the storage node. The commands to make this happen on a Unix server in ksh would be something like:

export RUSER=root
export RCMD=&quot;nsrjb -s backupserver -HEv&quot;
nsrexec -c storagenode

The procedure is very similar from an NT server, I'll leave you to work out the details.

You can also do a reset from the GUI. The NT GUI has a fairly obvious RESET button, on the Unix GUI it's a bit hidden away. You need to find the attribute of the jukebox called &quot;Operation&quot;, it's a hidden attribute so you need to do View-Details first. Pull down the Reset option and apply it. There are other boxes near to this to specify slots and other options. This is also the only place in the Unix GUI where you can do deposit and withdraw operations.

I don't quite know why Legato chose to make this so obscure in the Unix GUI. We're long overdue for some updates to the Unix GUI, which I use in preference to the NT GUI.

 

[605]

Hi DavinaTreiber,

this is terrific. I knew the NW only uses RPCs but i have never used it so far.
Let me add the windows stuff - the only difference are missing parenthesis:

C:\>SET RUSER=root

C:\>echo %RUSER%
root

C:\>SET RCMD=savegrp -v -G Default

C:\>echo %RCMD%
savegrp -v -G Default

C:\>nsrexec -c nw_server


Of course, nsrexecd also works cross-platform.

 

[does]

This is great!

Thank you very much.

 

[DavinaTreiber]

605's example shows another good application of this technique, typically used to start a group from a client.

I forgot to point out the security restriction, of course the same criteria used for backup commands, i.e. the command must reside in the same location as the save command, and must be named &quot;save....&quot; or &quot;nsr....&quot;. There is nothing to stop you making custom commands in this way, for example I use a command called nsrsjirdtag, simply created by adding a link (ln command) no prizes for guessing what that does.