Metrocluster IP replication types

[Jerry2]

Dear NetApp Gurus!

Let's take a deeper look on how to replication in MetroCluster IP configuration works. In my opinion, replication of storage data is performed over the iSCSI connections and replication of NVRAM data and storage data is being perfomed over high bandwith dedicated links using IP fabric. There is no FC fabric deployment. Am I right?

There is documention for Metrocluster IP

https://www.netapp.com/media/13481-tr4689.pdf

[SpindleNinja · Accepted solution]

Yep.    

Here's from the TR - "Replication in MetroCluster IP MetroCluster IP leverages direct attached storage, which eliminates the need for external SAS bridges to connect disks to the storage fabric. Each node in the disaster recovery group acts as a storage proxy or iSCSI target that exports its disks to the other nodes in the group. iSCSI (SCSI over TCP/IP) is the storage transport protocol for the IP fabric that allows the iSCSI initiator and targets to communicate over a TCP/IP fabric. Each node in the disaster recovery group accesses its remote storage through an iSCSI initiator that establishes an iSCSI session with a remote disaster recovery partner iSCSI target. The use of iSCSI and direct-attached storage also enables the use of systems that have internal disks. iSCSI allows the nodes to provide the disaster recovery partner node access to internal storage in addition to storage devices located in external disk shelves. MetroCluster has three planes of replication:

1. Configuration replication

2. NVRAM replication

3. Storage replication

https://www.netapp.com/pdf.html?item=/media/13481-tr4689pdf.pdf

There's two variations as well,  dark fiber / dedicated ISLs. as well as an option for Shared L2 ISL.  (page 27 in the TR)