Hey,

You cannot assign both SPA and SPB to one Storage resource on VNXe.

Because, in a dual-SP VNXe system, Storage resources are distributed between the two SPs; however, a storage resource is only assigned to one SP at a time. For example, a Shared Folder storage created on SPA will not be associated with SPB unless an SP failover occurs.

An SP fails over when it reboots, experiences a hardware or software failure, or a user places it in Service Mode. In this case, the storage resource is failed over to the peer SP. The surviving SP assumes ownership and begins servicing host I/O requests. When one SP is servicing all I/O requests, performance between hosts and the VNXe system can be degraded.

Below are FSN related information.

Fail Safe Networking

Fail Safe Networking (FSN) is configured on the VNXe storage systems by default. If one of the ports in the storage system fails, FSN will automatically re-route the I/O internally to the corresponding physical port on the peer storage processor.

For example, you can have port ‗eth2‘ on SPA plugged into switch A; and port ‗eth2‘ on SPB plugged into switch B (both of these switches must be on the same subnet). If a host application has data going over port ‗eth2‘ on SPA and switch A fails, FSN will route the I/O traffic over port ‗eth2‘ on SPB using switch B. Fail Safe Networking therefore helps in providing switch level redundancy in your environment.

And you can get more information from EMC VNXe High Availability Overview .

And from an operational perspective can also explain the reason

As you may know if you want to create a "Storage Resource"(iSCSI and Share Forder etc.), you need to create "iSCSI Server/Share Folder Server" first. When you create "iSCSI Server/Share Folder Server", you can only specify one Storage Processor(SP).

Then, when you create Storage resource(iSCSI and Share Forder etc.), you need to select a "Storage Server" which is just created iSCSI Server.

So, you cannot assign both SPA and SPB to one Storage resource on VNXe.

Hi Leo,

Thank you so much for the quick response!

But i did not understand Fail Safe Networking fully.

For example, i have created two data stores, datastoreA and datastoreB. SPA is assigned to datastotreA and SPB is assigned to datastoreB.

Assume Virtual Machine A is assigned to datastoreA and all its files are in datastoreA. As per FSN, if SPA fails, SPB will take care of the traffic. But here SPB is assigned to datastoreB. In this scenario, how does Virtual Machine A can retrieve its files from datastoreA.

Regards,

Bineesh

Hi Bineesh,

As per Fail Safe Network concept, It provides port, cable or switch redundancy. This is not failover of SPs. If any port, cable or switch is failed, services will still run on owning SP but traffic gets re-routed from other SP. It uses a internal connectivity between SP for this.

This is transparent to host. If network switch is configured properly, host will not loose the access.

Example:

- SPA eth2 is configured with 10.10.10.1 and SPB eth2 is configured with 10.10.10.2.

- In normal situation traffic will be routed from owning SP

- If eth2 on SPA is faulted, all services will continue to run on owning SP and only traffic will be internally routed SPB eth2. Host will not loose access until, SPB eth2 has access to virtual machines.

- Datastore owned by servers will always stick with it and will be accessible from which ever SP it is on. Saying that, we are creating datastores on servers (iSCSI, NFS, CIFS) which is associated to individual SP.

In case of reboot, servers failover from one SP to other.

Regards,

Swapnil

Hi Swapnil,

Thank you so much for the update.

I am bit confused. As per your example if eth2 on SPA (10.10.10.1)is faulted, traffic will be taken care of by SPB eth2(10.10.10.2).Traffic flow from storage to Virtual machine is ok.
How does traffic flow from Virtual Machine to storage work?

Virtual Machine only knows SPA - 10.10.10.1. How does Virtual Machine redirect traffic to SPB-10.10.10.2 if there is any failure on eth2 SPA. Do we d to have any additional configuration on Virtual Machine side?

Thanks in advance.

Regards,

Bineesh

Hi Bineesh,

Switch will take care of this. When eth2 on SPA is down and traffic re-routes from SPB eth2, now a Layer-2 managed switch should be able to detect the path failure and will be able to resume communication from SPB eth2 port.

Each ethernet port can have multiple IP addresses and will do a failover if SP reboots. Also, FSN will work if one port is down. Switch takes the initiative here and reconnect the session back from other port.

Regards,

Swapnil