Hi there, we are currently trying to setup a new MCC-IP with 2 x C800 clusters (4 nodes). We have cabled "everything" but the e0M into the Nexus C9336C switches. The switches are configured using the RCF tool with configuration for two Metroclusters and SAS and Ethernet storage. So all ports but 27+28 are configured. (we are of cause not using all the ports right now).
The Intercluster links can be either dedicated interfaces or shared with data-interfaces which will direct intercluster traffic through the frontend switches... is it not possible to use some of the backend ports for this traffic?
Is it not supported to do it this way? I gues it wouldn't be that big a deal to use the two unused ports 27+28 for this? Or should we just stop worrying and share the traffic with the data (iscsi) ports? (we are talking 100G links on everything, even ISLs on both front- and backend)... also if we choose to configure shared IC links, is it as easy as just moving the lifs to other ports if we wanted to later on?
#Questions not found the manual when configuring a MCC-IP...
deep viper
deep viper
....if we were to go for shared intercluster links, is it supported to create them on a separate VLAN?
unborn maple
Intercluster is always through a frontend network. There is almost no traffic going through that network anyways ...
Putting them on the backend switches is not supported because you're not supposed to edit the RCF file in this way
deep viper
Yes I found the details about the RCF files, so I guess you are right we better leave it alone 😉 We will create a seperate VLAN for the intercluster links, which I cannot see mentioned should be a problem...
Someone at NetApp should take a closer look at the documentation of the setup, because a lot of it doesn't follow what actually happens... 🙂 I think they forgot to update it...