Smith, Richard wrote:
>
>         I would also be interested in this.  We are on an OS/390 server
> though, and our reason is that the UNIX guys want to handle their own
> stuff.  Any info would be appreciated.
>
> Rick Smith
> Maritz, Inc.
> Storage & Security Administration
> smithrr AT maritz DOT com
> (314) 827-1584
>
> > ----------
> > From:         Trevor Foley[SMTP:Trevor.Foley AT BANKERSTRUST.COM DOT AU]
> > Sent:         Monday, March 30, 1998 10:59 PM
> > To:   ADSM-L AT VM.MARIST DOT EDU
> > Subject:      Multiple ADSM servers on one AIX system
> >
> > Hi,
> >
> > We are looking at running multiple instances of the ADSM server on one
> > AIX system (we have a small number of clients that, for security
> > reasons, we don't want to be on the main ADSM server). Both ADSM
> > server
> > instances will be the same version (currently 3.1.0.2), and our AIX
> > version is 4.2.1.
> >
> > Has anybody else done this? What tricks do you have to do to keep the
> > two servers from accessing each others data? We are already running
> > our
> > first ADSM server with a default directory of other than
> > /usr/lpp/adsmserv/bin, so we are part of the way there.
> >
> > Any suggestions would be appreciated.
> >
> >
> > Trevor
> >

Oops, hit the send button too quickly...

We have done this for our HA clusters.  We also ported the HA scheme to
other "single" servers to comply with conventions.

We build a filesystem that holds all of the server information:
dsmserv.opt, volhist data, devconfig data, recovery plans, build
scripts, etc (we also mirror this information offsite for DR purposes).
We name our servers in ascending numerical order and build filesystems
which coorespond to that order, i.e., server01 might have a filesystem
called /adsm1-home.  We gen the database from that directory.  We have
the client options file and an environment file which sets all
DSMxxxxxxx environ variables appropriate to that server in the server's
filesystem.  dsm.sys resides in the common binary install location but
always has every server listed.

There really aren't any tricks other than following a rational
convention.

David H.
dmhendri AT fedex DOT com

--
Man usually avoids attributing cleverness to somebody else --
Man usually avoids attributing cleverness to somebody else --
unless it is an enemy.
                -- A. Einstein