Does anyone have any experience setting up a group of servers for high reliability/availability?

We are planning to deploy a WebObjects application but need to design a configuration that will give us a VERY high level of availability. Our current thoughts are to have three machines:

 1. Xserve G5 with mirrored Hard drives running
    - MacOS 10.3.x
    - mySQL  (Replication - master)
    - WebObjects (HTTP Adapter)
    - WebObjects Application instance(s)

 2. Xserve G5 with mirrored Hard drives running
    - MacOS 10.3.x
    - mySQL  (Replication - slave)

 3. Xserve G5 cluster node running:
    - MacOS 10.3.x
    - WebObjects Application instance(s)

If machine 1 fails then machine 2 will be configured with IP Failover to become the mySQL master and WebObjects would be launched with the adapter and our WebObjects application instances.

One issue we have found is that when machine 1 reboots, it would failover to machine 2 momentarily. MySQL would be configured to automatically send its data back to machine 1 when it comes back up. For a simple reboot this seems unreasonable, it takes about 20 minutes to send the data back (tables would have to be locked).

If machine 2 fails then nothing happens.

If machine 3 fails then the WO Adapter would no longer delegate requests to it.

As our needs increase we would add additional cluster nodes running WO Application instances.

Eventually we will migrate to clustering, mySQL v4.1.

Any thoughts, comments?

Thanks,

kib
--
Klaus Berkling
Systems Administrator/Engineer
DynEd International, Inc.


-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]



Reply via email to