1

I am very new to storage but we are having trouble with performance on our SAN. While troubleshooting I found that one of our switches in the fabric (1 HP C7000 with 2 flexbric interfaces connected to 2 SAN Switches connected to 2 HP 3Par SAN nodes) had pretty much no traffic.

I was wondering if that kind of situation was "expected" (maybe it's using 1 switch as primary and the second as failover) or if it could indicate some problem.

When running a single "porterrshow" on the Brocades, or "portperfshow", I can see a lot of stats on Fabric_A but almost nothing on Fabric_B.

Is it normal to have traffic on just one interface, and not load balanced across interfaces?

Let me know if you need any details on the setup.

kasperd
  • 29,894
  • 16
  • 72
  • 122
user3249382
  • 11
  • 2
  • You need to configure the hosts to use paths on both fabrics. – Basil Oct 24 '18 at 13:03
  • thank you Basil, so the problem might be on the hosts side ? i will check there (as we have the exact same setup on another site and it works perfectly fine there) – user3249382 Oct 24 '18 at 14:35

1 Answers1

0

we indeed had a wrong storage policy configured on the VMware/ESX side. We reconfigured it from MRU to RR (round robin) and can now see all paths being used and traffic on every switch

user3249382
  • 11
  • 2