Hi,

We have two Baan Servers running informix 9.40FC5 on Sun Solaris 9.x.
The both servers are connected to SAN on shared Storage Array 6140x and the database size is nearly 300Gb each server. Both Server are having different baan companies numbers.

We plan to configure two Instances of Informix on each Server and configure it as failover. The 2nd instance on each server will have pointer to the 1st instance of other Server.

Server (A) Server (B)
Instance#1 (Online) Instance#1(Online)
Instance#2 (Offline) Instance#2(Offline)

If Server (A) goes down, the 2nd instance on Server (B) will be put online and the users can now login into Server (B) and do their work.

Would appreciate if anyone can share his views on this setup?
- Technically possible but are there any issues to it?
- What are downside to it?
- Has anyone setup like this before?

Regards
Muhammad Usman
musman@iffco.com

- Has anyone setup like this before?

Maybe I can give you a hint/an idea... .

First (just for the protocol :) ):
I'm not our Unix guy
I'm not a DBA

I'm just an interested Baan Tools Consultant/Programmer/System developer

System description:
1 SUN for "heavy users", jobs, ...
1 SUN for database (~ 1TB), database Oracle 8.1.7.4 (64 bit),

Both maschines are able to power the database and the Baan Jobs (slower then normaly, but it works)

Both maschines are in a HA environment. Switch time: 10 Minutes (we had a switch today because of hardware trouble).

HA licence manager on the maschines for all servers

4 (smaler) SUNs as application servers with a (two because of backup) loadbalancer in front. The users connecting to the LB (normaly :rolleyes: )

The whole Baan system except of $BSE/log and shared memory control files are shared on a SAN (by NFS, SAN capacity alltogether ~7 TB (?); the "disks" of our Windows server(s) (not Baan; ADS, Exchange, fileserver only) are also on the SAN). shmvalues and $BSE/log are local on every maschine.

One server for developing/testing (DB (~700 GB) and files are also on the SAN, licence from the cluster)

hope this helps

Regards
csecgn