WM redundancy......

[jriopel]

Does anyone have in place some clustering or mirrored file systems to provide 100% uptime for WorkManager? I'm sure i cant be the first person in the world to have this need.
I have 2 identical servers + disk arrays and would like to implement something to allow a failover to the second server if one should fail. ORACLE's RAC and simple mirrored file systems are somethings im thinking about. Both oracle and area1 are two seperate volumes so im wondering if i need to do one or both. Also, HP's MC service guard is not supported on the hardware i have so it would have to be somthing else. GDBMgr will not work because i need the updates to area1 and the db instantaneously. Any info would be greatly appreciated.

[ca32316]

Hi
This does not answer your question directly, but may be something to think about if 100% uptime is not mandatory. We put wm and all related oracle stuff on a NFS file server. Network Appliance Corp (NAC) to be specific. NAC has been certified to run oracle. They can provide failover capability between 2 of their boxes. We use a DNS alias to reference our server. With all of the data on a remote system and a alias for the server name, the wm server is extremely portable. If one server goes down, update a few oracle files on s new server, change the DNS alias to the new server, check the mountpoints on the new server, fix any oracle crash related issues and off you go.
This is by no means failover, but you can have a new server up with all the same stuff in a few minutes depending on how much damage was done to the oracle tables during the crash.
NAC also provides Snapshots that can be automated to keep images of the file server at specified intervals. We keep rolling hourly snapshots during working hours which consume ~ an additional 15% of disk space.