Re: FailSafe

FailSafe takes advantage of Cluster Server by putting the database, and listener, on the clustered storage device using a third IP address. An instance is created on one server and FailSafe propagates that instance/service to the second server in the failover set. Only one server can have an active instance and, through the functionality of Cluster Server, NT allows Oracle to monitor and manage the live instance and also know when it is down, to facilitate a fairly quick failover to the remaining working server. Since all database files are on the storage device (init.ora, control files, database files, listener, etc.) the transfer of 'responsibility', shall we say, in the event of a failure is facilitated. Also, since the listener address does not change should server 2 take over clients do not need to change their tnsnames.ora file. Since a NAS device on NT allows only one server or process to lock a file or set of files this ensures that only one database server is up and running at any given time. The second server cannot take control of the database files until the first server 'dies' and relenquishes control. Of course, if the RAID array (or other storage device) fails, FailSafe also fails, one drawback to using FailSafe. (Of course, in a Parallel Server environment the same situation occurs with a RAID/NAS failure.)

I hope this provides some clarification.

sjefke12_at_hotmail.com (Foggy) wrote in message news:<f78e018a.0206202227.317521e2_at_posting.google.com>...
> Hello,
>
> I have a question about the benefits of Oracle FailSafe. I have read
> the FailSafe Concepts Manual, but it is still not clear to me what
> FailSafe adds to the Microsoft Cluster Server functionality.
> Could you point out what FailSafe adds to Cluster Server?
>
> Oracle 8.1.7 on Win2K
>
> Regards,
>
> Johan
Received on Fri Jun 21 2002 - 08:12:22 CDT