Good evening. I assume many of us have an idea about disaster recovery and I need suggestions for implementing one. My client have a production Env(Clustered) and a DR env(non clustered). While simulating, they have simply copied the exact DB of IS and MWS and replicated to DR database. I faced some problems due to MWS cluster.xml and host specific config files. But Integration Server didn’t face any problem.
Does anyone has best practices of doing it, also broker data. Is it possible to import storage files or such data from one server to another independent server. Was looking into it for quite a while, but didn’t get much info.
We have 2 identical, clustered environments and network based storage.
The Storage and the DB (Oracle Dataguard) are permanently replicated from the main environment to the fail-over environment.
For the case that the main environment will be unavailable, the fail-over environment starts the services and instances based on the latest replicated storage data and DB data using the same logical IP-Adresses and DNS-Aliases as the original environment.
Hi Holger
How do you manage to keep same logical IP-Addresses as its not possible to have same IP addresses among different data centers. I am trying to set up My webMethods DR and we only have task engine running on My webMethods which works as a back end workflow engine without business users logging in to My webMethods
Hi Holger
Thanks for your reply. When you say logical IP address is assigned to services, you are referring to which services ? Can you elaborate more on “The DataCenter is spanning over two datacenters sharing the same IP adresses.”
servive in this case means a resource service as of Oracle Solaris Cluster control, bundling certain resource groups which contain mounted filesystems, ip adresses and application config data.
So when starting such service on either side of the DR DataCenter, the application will always look fron outside for our clients when they connect to it.
Only difference is when we connect to our partners they will see that we are sending from a different IP adress then before.
The DR DataCenter is spanning over two DataCenter locations which are several kilometers apart from each other.
My company names these Twin-Core DataCenters as these look like one big DataCenter from a higer point of view.
Hi Holger
Thanks for your reply. We are a Windows shop so don’t have the liberty of using Oracle Solaris Cluster control. I have not seen any limitation regarding IP address so was thinking to replicate the file system and database schema at the same time into DR environment at regular intervals. Basically DR environment would be down like a traditional DR and then when DR is initiated following steps would be executed
DNS is switched between main environment and DR environment
db.xml is restored from DR environment as DR database credentials would be different