We have a two-node AFF700 cluster running 9.3P4. We have two ipspaces defined, one for general traffic including presenting data and replication to backup NetApp devices. The second ipspace is dedicated to interfacing with StorageGrid (SG) and specifically Frabic Pools. We have dedicated physical ports allocated to the SG ipspace. We have broadcast-domains, lifs, etc in the appropriate ipspaces.
We created the separate ipspaces so we could segregate the intercluster lifs used for client data access/replication from those used for StorageGrid. The goal was to achieve the full-mesh connectivity requirement.
The issue I have is the network throughput between the AFF cluster and the StorageGrid is terrible, 3-5 Mbps on a 1Gbps network. We spent too much time confirming that the issue is not related to our enterprise network.
One observation made this afternoon is that both ipspaces have lif's configured to access the same vlan (113). Can anybody tell me if there is a problem configuring lifs in two separate ipspaces to access the same common vlan?
If two ipspaces connecting to the same vlan won't work, please tell me why.
if it's still smell like a network issue. check ports stats on the nodes and switches, and if still in question - gather a packet trace on the filer and client and try to find issues. here's one example of an issue that likely to hit 1GB interfaces and can be picked from the port stats and packet trace (but there are a few more i can think of):