trying to purchase a highly reliable ORACLE server configuration...
Sorry for wasting everyone's time but after wading through reams of
product literature and some very vague salesmen I thought I would try my
luck here..
We are looking to purchase a system that will have no single point of
failure for running our enterprise database. We would like zero down
time (or close enough to zero, like 1 minute).
We found one solution, the IBM HACMP 6000 system, (High Availability
Clustered Multi Processing) which is 2 RS6000 servers connected together
with a heartbeat, both accessing the same RAID bank. Oracle's Parallel
Server would be running on this system. If one server goes down, the
other server picks up the users and continues on.
We would like to see whether or not we can get a similarly highly
available system using perhaps 2 traditional UNIX servers, 2 banks of
disk and keeping the two instances of the database in sync through
replication services or perhaps a transaction monitor and achieve
automatic fallback and re-synching (in the case of the HACMP, the
fallback is done through IBM's software and there is no synching needed since
there is one RAID), and do so with as close to 0 down time as possible.
It seems at this point that with replication services, we do not achieve
automatic takeover and there is no automatic re-synching of databases
when the problem is over and the database is brought up. Although I
have been told that using a transaction monitor this capability could be
implemented, I do not know whether ORACLE has implemented this
capability.
We have no specific inclination towards any hardware platform. Cost,
although always an issue, is not our greatest concern.
Any help would be much appreciated.
Thanks,
Deniz Akkus
Received on Thu Sep 15 1994 - 17:52:04 CEST
Original text of this message