Cluster failover

[Jawloms]

Hi All,

We are looking at implementing a cluster failover system here so that if the SIMS server dies, another will just kick in and take over. Has anyone else done this? If you have, could you either reply to this post or e-mail me directly (shawkins@orwellhigh.com) as I have only just started looking in to this and it looks a little tricky (to say the least!!)

Thanks for any help

Stuart

[fafster]

How about running the SIMS server virtually? There'd be a delay between the time it took to move the virtual machine to another host, but it would be less problematic than trying to set up a cluster.

In fact, I'm sure you could store the SIMS database on a DFS share (It's MS SQL express, I think), and run the server in the virtual machine. Then store multiple copies of the virtual machine on various computers and simply switch them on if needed.

[Jawloms]

Thanks for the idea, the trouble is the head wants it so that we don't have to take time out to switch anything over. He wants it so that if the server dies for any reason, the other one just starts working instead and the users don't notice anything has happened.

Stuart

[SYNACK]

I think that you will need to get into SQL database replication between the two servers as clustering by itself will probably not cut it when it comes to the continuity of data. Same thing with DFS the file would be considered open and therefore not replicated until the service closed the connection to it.

[localzuk]

The problem here would be that the sims client uses the connect.ini file to connect to the server, and this will have the DNS name of the server hard coded into it will it not?

[SYNACK]

If you are using a combination of clustering and direct SQL replication the servers should show us as a single entity that is referenced by the connect.ini file.

[Oops_my_bad]

Clustering? Failover?

Come on.. this is SIMS we're talking about here xD