Title: Safe Database Environment
Hello Hussain
 
We are working with oracle 8.1.6 on wintel machines.
 
We had the same problem and we solved it using advance replication - synchronous.
We have scripts ready that drop the replication on each computer if one become unavailable.
The databases have the same sid.
 
So if one computer goes down we do the following:
 
1) Drop replication on the remaining server.
2) If main server is down:
    remove the backup server from the domain
    rename the backup server name and tcp/ip address to the main server name and tcp/ip address.
    start up the backup server as main server.
    IMPORTANT - run a script that update the values of sequences, since they are not replicated.
3) Start the application again.
 
Elapsed time about 5-10 minutes and ALL the data is in.
 
Yechiel Adar
Mehish
----- Original Message -----
Sent: Thursday, November 21, 2002 10:13 AM
Subject: Safe Database Environment

Hi all,

Can any of you suggest the configuration for a safe, fault proof, and minimum downtime Database environment? (In terms of how many servers, RAID or disk ARRAY or any combination. A minimum configuration is what I am looking for).

Right now we have one main server (Compaq proliant) and one standby database. But since the log switch takes place after about 10-15 minutes in the peak time and 30-40 minutes in not so peak time, plus our environment is 24x7, so if anything goes wrong with the production DB, it will take sometime to get back live and then some further time to catch up or redoing those entries which were lost because of no log witch. Hence even a downtime of few minutes brings the work to a halt plus any loss of data in between the archiving time is costly too.

So I want to suggest to our management a solution, which is as close to being fool-proof, if one server is down the other system can take its place in matters of minutes or seconds, and better if it is transparent to the user, or close.

Thanks & regards,
Hussain

Reply via email to