Hi
You may need to tell us a little bit more about your configuration especially your fabric/fibre infrastructure, and host configuration.
Ideally with a NetApp cluster and FCP hosts you'd have two HBAs in the host each connected to a separate fabric switch, each switch would then have a connection to each NetApp head. This provides redundancy in case of the failure of an HBA, a fabric switch or a filer.
Assuming the lun is on Filer1, in normal working the host will see 4 paths to it.
HBA1 to SW1 to the Filer1 directly
HBA2 to SW2 to the Filer1 directly
HBA1 to SW1 to the Filer1 through the interconnect with Filer2
HBA2 to SW2 to the Filer1 through the interconnect with Filer2
There a few things that could be happening in your case:
1. Things are connected wrongly or the fabric switch are not configured properly and the only path the host is seeing to its lun is via the filer interconnect.
2. Some traffic is going through interconnect when the preferred path is busy and you've nothing much to worry about.
3. The host system is using all the FC paths with equal priority, you may be able to change this behavior, depends on the O/S and HBA driver.
A few commands to run the filer command to determine the situation.
lun stats -o will give a an outlook a bit like this
/vol/yourvol/yourlun.lun (13 days, 23 hours, 46 minutes, 46 seconds)
Read (kbytes) Write (kbytes) Read Ops Write Ops Other Ops QFulls Partner Ops Partner KBytes
2574820375 5211323117 9785526 208670967 112320 0 30619 16
Do a bit of maths, compare column 8 to the sum of column 1 and column 2. If it's more than 25% I say you've a serious problem.
lun stats -z reset the counters to zero.
Good luck
