High Availability - Windows

My company is currently in the process of re-platforming its hardware infrastructure for E1, provisioning greater reliability and availability. A requirement is to implement an active/active MSCS high availability failover solution for our enterprise servers consisting of separate UBE and business function servers. I’m reaching out to the JDE community help on this configuration - has anyone considered or successfully implemented this solution?

Any help will be greatly appreciated….

Regards,

[ QUOTE ]
My company is currently in the process of re-platforming its hardware infrastructure for E1, provisioning greater reliability and availability. A requirement is to implement an active/active MSCS high availability failover solution for our enterprise servers consisting of separate UBE and business function servers. I’m reaching out to the JDE community help on this configuration - has anyone considered or successfully implemented this solution?


[/ QUOTE ]

Sydney,

Yes, someone has. We have had a similiar setup for over five years. We are actually in the middle of round two, swapping out all of the servers with newer, bigger and better.

Enterprise server - two eight way HP / Intel servers
Storage on a very large SAN
Three seperate instances of MS SQL on the boxes. One for JDE Prod, One for JDE test, One for Cognos. Each instance of SQL is managed in it's own cluster group to allow them to move back and forth across the cluster.

The cluster is split between two different buildings, a mile apart, to provide for disaster recovery. I have three other batch servers to support batch jobs, webmethods interfaces and JDE web client. Those are also split between the buildings. We have fourteen terminal servers, seven in each datacenter.

Gregg Larkin
Praxair North American System Admin
JDE CNC and Security, Websphere, Tidal, Princeton Softech

I have set up at my company, (besides our legacy JDE system) an array of 4 JDE EnterpriseOne apps servers (soon to be put on a load balancer (LB)) 6 JAS servers (already loadbalanced) , one back-end database and, of course, the ever-present deployment server. With this configuration we can take any JAS server off the LB and reboot it or do work on it. Likewise, we'll soon do the same with the apps servers, meaning that (in mid-day or anytime 24/7) we can take a couple apps servers off the production to deploy packages to them, and then add them back onto the LB. This is a true horizontally scaled JDE ERP system. Instead of apps security server PD01 or PD02, etc, all OCMs point to PD00, which in turn crosses over to the load-balanced apps servers. Same with the JAS servers. Also, all apps servers, in their JDE.ini, share the same printqueue folder and we've sorted out the print job table with triggers to handle changes in how this works. JDE published a VERY limited doc on this in June 2003. They told us, recently, however, that this solution is NO LONGER supported without their people coming on site to do the set up. All this runs in the Windows and SQL-2000 world, but we're planning to move off to Linux and Oracle's so-called "Red Stack" solution with the production database also running on Oracle 10g. When we move over to a horizontally scaled Linux-based Oracle Transaction Server and Oracle database, we should have no trouble scaling out to 10,000 users.

[ QUOTE ]
[ QUOTE ]
My company is currently in the process of re-platforming its hardware infrastructure for E1, provisioning greater reliability and availability. A requirement is to implement an active/active MSCS high availability failover solution for our enterprise servers consisting of separate UBE and business function servers. I’m reaching out to the JDE community help on this configuration - has anyone considered or successfully implemented this solution?


[/ QUOTE ]

Sydney,

Yes, someone has. We have had a similiar setup for over five years. We are actually in the middle of round two, swapping out all of the servers with newer, bigger and better.

Enterprise server - two eight way HP / Intel servers
Storage on a very large SAN
Three seperate instances of MS SQL on the boxes. One for JDE Prod, One for JDE test, One for Cognos. Each instance of SQL is managed in it's own cluster group to allow them to move back and forth across the cluster.

The cluster is split between two different buildings, a mile apart, to provide for disaster recovery. I have three other batch servers to support batch jobs, webmethods interfaces and JDE web client. Those are also split between the buildings. We have fourteen terminal servers, seven in each datacenter.

Gregg Larkin
Praxair North American System Admin
JDE CNC and Security, Websphere, Tidal, Princeton Softech

[/ QUOTE ]

Wow, i would love to see your datacenter.

[ QUOTE ]

Wow, i would love to see your datacenter.

[/ QUOTE ]

Madmax, That's only a small part of one of our datacenters. That's in the Danbury Ct center. We have major Data Center's in Buffalo, NY; Milan, Italy; Rio De Janario, Brazil; and Singapore. We have a bunch of minor data centers scattered around the world.

Gregg