ONTAP Hardware
ONTAP Hardware
Hi
weve got a C250 in the setup phase currently and by looking at the below KB the HA and Interconnect cabling is shared on the e0c and e0d
HA interconnect procedure guide for ASA, AFF and FAS platforms - NetApp Knowledge Base
i was told by the NetApp presale that there is a work around by seperating the HA and IC ports, we have an additional FC mezzanine card on the C250 for both nodes so i wanted to use 2 ports from that to seperate the HA and IC.
I asked how to set it up and was told the "move the port" using the below KB
How to move a cluster port from one port to another in ONTAP? - NetApp Knowledge Base
wouldnt this defeat the purpose of sepearting? i wanted to use e0c/e0d for HA and e2a and e2c port for cluster interconnect.
im really confused on how to set it up, is the above procedure correct? is there a specific procedure to set up "this port is for cluster interconnect" or does moving or adding port automatically recognize the port is used for HA and IC?
TIA
Solved! See The Solution
Hi there!
If you just have an HA pair, there is extremely limited utility from splitting HA and CI onto different ports. A single link failure of the combined port won't take the system out, nor will splitting them onto different NICs save you if both links of a pair fail. The benefit is that it enables somewhat more speedy CLAM takeover vs slower HA takeover in case of a node failure, but it's an extremely uncommon issue.
We do support it being done if the qualification is approved by our engineering teams. Your NetApp account executive will need to lodge a request with them for that qualification. If they need guidance on the process, they can contact me internally.
Hope this helps!
Hi there!
If you just have an HA pair, there is extremely limited utility from splitting HA and CI onto different ports. A single link failure of the combined port won't take the system out, nor will splitting them onto different NICs save you if both links of a pair fail. The benefit is that it enables somewhat more speedy CLAM takeover vs slower HA takeover in case of a node failure, but it's an extremely uncommon issue.
We do support it being done if the qualification is approved by our engineering teams. Your NetApp account executive will need to lodge a request with them for that qualification. If they need guidance on the process, they can contact me internally.
Hope this helps!
hello! i did ask the netapp sales and they told us that it would be supported, after that they sent us the KB in order to move the port. Upon speaking with the netapp staff again earlier, i was told that even if we move the Intercluster connect port to e2a and e2c, the HA port will still reside in e0c and e0d, hence being able to create the seperated config. i understand the possibility of 2 HA cabling goes down is slim (knock on wood) but seen as we have spare ports from the mezzanine card, we wanted to sepearete the ports
You require an approved feature product variation request (FPVR) from NetApp to support this use-case. The instructions are included when the FPVR is approved. Your NetApp sales representative needs to file this request.
BTW, you mentioned "FC mezzanine"... I think you are referring to the X1152 10GbE/25GbE 4-port module, right? Just checking... "FC" usually means Fibre Channel to many people.