On Mon, 14 Apr 2003 16:30:28 -0400, Gaddy <xy.0815 AT GMX DOT NET> wrote:

>Hi all,
>
>there has been a thread about DDS
>and it's "hidden" limitations (scalability).
>But some time has gone by, ...
>
>We have an installation with IBM3584, 6xLTO-1,
>and do backups approx. 1500 GB daily (max.)
>using NetWorker 6.2 on W2K.

Take care with version 6.2 (I will use 6.1.3, if posible)

>So we've decided to use SANnodes (all W2K) for
>- applications, because they do daily fulls
>  (notes, oracle, ...)
>- some bigger file servers, because restore
>  time has to be as short as possible
>
>At the momenent we have 6 DRIVES and 42 DEVICES,
>which means:
>6 SANnodes share all the 6 drives
>2 SANnodes share only 3 each (they only have 1 HBA)

Don't forget to reduce the "Idle Device Timeout" in jukebox window.

>With this configuration we see MASSIVE RPC problems
>nsrmmd timeouts, nsrmon exits and all the like.
>(not to mention the silly "jbconfig" sessions)

You can use "jbconfig < file_cfg", where file_cfg is a file containig the
jbconfig dialog, so if you need to reconfigure the jukebox you can do it
easier.

>So my questions are:
>Has someone good/bad experience with tuning the
>nsrmmd settings?
>nsrmmd polling interval = 3
>nsrmmd restart interval = 2
>nsrmmd control interval = 5
>What will happen if I increase polling+control interval?

Use persisten binding. W2K could change, and it does, the names of device
tapes (\\.\TapeX), that confuse backup software.

Regards,
Gerardo.

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