Cirrus
MVPF5Team I have answered these as best as I can but the DCD documentation is a bit lacking.
1Q) If we have multiple DCD's, would the statistical data replicate completely between them or in parts?
1A) In a DCD cluster configuration is replicated between themselves but stored information they receive is typically to a storage server which you map to with the DCDs.
2Q) If a logging profile has 2 DCD IP's configured under server address, would only the first DCD receive all the data until it goes down n then the second DCD gets the data?
2A) I believe these 2 IPs would be 2 different cluster IPs if you had two seperate clusters for redundancy, such as primary site and secondary site and I believe logging would occur to both simultaniously so that each site would have this information in the event of a complete loss of data at one of the sites.
3Q) If the data is replicated between DCD's then what is the need to load balance the BIG-IP traffic to DCD?
3A) The data itself is mapped to a storage server/device and is pulled by the cluster when the CM retrieves this data. I believe the only part that replicates between DCDs is their own configuration and not logging information.
4Q) If we are load balancing the BIG-IP data to DCD, considering requests are going into diff DCD, what happens.
4A) This information is all stored at a centralized storage location for the cluster so you should not have any duplicate logs or anything that the DCDs retreive themselves.
5Q) If one of the dcd goes down? will the BIG-IQ be able to pull the stats?
5A) If you have the DCDs in a cluster it would require a failure of x number of DCDs in the cluster before the CM couldn't retreive this information. Because you have a storage server/device that the DCDs map to you should be able to restore a DCD and then the mapped drive will still be in place so the data is retreivable.
It is possible that I might have interpreted this information from the DCD documentation so if anyone else has input please do.