Hi Stuart,

I need to setup an ADSM system that will fail over to another RS/6000 if
there is a problem with primary system.

The system as it has been installed (not by me!) has the database and
recovery log volume group failing over - which is obviously required.

What do you mean by "failing over" ? On shared disks ?

Not only the database and logs should be on shared disks, but
also disk pools and configuration files (dsmserv.opt, volhist
and devconfig files).

Two considerations :

First, disks should be arranged on fault-tolerant schemes, like
mirroring or RAID-5. ADSM allows mirroring the DB and Log. Be
sure the mirrors are placed on different disks, or a single
disk failure could bring down your HA solution.

In this case it is easier to use AIX mirroring (what leads
to another discussion of ADSM or system mirroring, but this
is not the main issue of this message).

No need to say that tape drives and libraries should be
shared too, of course.

Second is a particular opinion that although ADSM is
installed on /usr/lpp/adsmserv, you should never
keep your configuration files and even your ADSM
startup directory on ROOTVG, because if you needed
to restore from mksysb tape you would be erasing
important application configuration.

But
I can see no way of telling ADSM on the backup system to use this database
and recovery log - any ideas.

The only way I can think of doing this is to fail over the binaries as well
as the database and essentially fail everything over.

Don4t fail over the binaries. Let them rest on ROOTVG.

Create a separate VG for ADSM on the shared disks. Place
such things as DB files, Recovery log files and
Disk pools on different file systems in this shared VG.

Also in this VG, create a file system for ADSM startup.
It doesn4t have to be very big, just enough to hold :

volhist file
devconfig file
dsmserv.opt
adsmstart script
dsmserv.dsk

Change adsmstart script to start ADSM from this
directory, not from /usr/lpp/adsmserv/bin. ADSM
is bound to the directory it is started and
searches for dsmserv.opt file in it.

When you setup HACMP, remove ADSM entry from
/etc/inittab and place it on the Resource
Group startup scripts, so that the cluster
will fully control resources.

Also, to use the Web Admin client, create
symbolic links from *.GIF and *.class files in
the /usr/lpp/adsmserv/bin directory.

Be sure ADSM is installed on the failover
box, with same patches applied.

This configuration might do.

As an aside how does ADSM know where to look for its database volumes?

From dsmserv.dsk file.


best regards,
Rodrigo

______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com