1

I read this question. I have only 2 VMware esx with 2x1gbs interfaces each and netapp with 4x10gbs interfaces. It is configured auto negociate with the switch. But the switch drops frames.

Do I need enable full Control Flow between all interfaces ? Or is there any better solution.

For now we put limitation on the netapp bandwith on each datastore.

1 Answer 1

3

Flow control needs to be active on all host and switch ports to work.

The best solution would be to upgrade the host ports. With flow control active, you may experience head-of-line blocking when connecting an additional 10G host interface: when a saturated 1G link causes pause frames, the 10G storage port throttles down even if the 10G host port is waiting for data. This is a general design problem with Ethernet flow control.

Flow control is normally active on all ESXi network ports unless it is deactivated by using esxcli. The exact configuration depends on the type of NICs used. You can find details in https://kb.vmware.com/s/article/1013413 Essentially, you just have to make sure the switch and the storage array use pause frames as well - FC is most often deactivated by default on managed switches.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .