We are looking at implementing DECSAFE on some boxes where I work. What
we have currently is an 8200 with DU 3.2D connected to a storageworks
controller and a 2100 running DU 3.2c running on it's own internal
disks. We wish to be able to failover from the 8200 to the 2100 if the
8200 unit failed.
After getting a quotation from DEC, we realise that we need some extra
hardware, specifically some various cables and scsi controllers. It
also looks as though to implement the DECSAFE that we will need to
update both boxes to DU 3.2G. Due to some Oracle Financials requirement
we cannot upgrade beyond that point (I think there is some combination
that is not supported on 4.0).
After looking at the whole situation, it appears that the decsafe may be
a bit of overkill for our requirements, as a manual failover would be
adequate, and our users are going to lose their connections to the
machine if it crashes anyway. So, my question is this : Can anyone see
any reason why if the 8200 crashes/is powered down etc that we can't
change the IP address on the 2100, mount the disk from the storage works
and continue running in some limited capacity?
Also, we are not able to get much info from dec about exactly how
decsafe works and how we would go about setting it up. We are getting
quotes for quite large amounts of money for 'setting up a database
instance in decsafe'. Exactly what are they going to be setting up,
scripts, c code, etc? We feel a bit hesitant recommending to our
management to buy a product if all that we know about is that we which
we gleaned from a few marketing glossies.
Anyway, if you got this far, thanks. I will summarise to the list if
appropriate.
Brett.
--
Brett Rees
Unix Systems Administrator
email: breree_at_ozemail.com.au (preferred), breree_at_accsoft.com.au
phone: +61-414-210-194
home page: http://www.ozemail.com.au/~breree/ (old and outdated ...)
Received on Sat Nov 23 1996 - 12:40:03 NZDT