2011-10-11 01:31 AM
Im new to Netapp and have a FAS3210 2 x heads, dual Brocade 510's and using MPIO & Netapp multipathing. We are using FCP with LUNS to a MS 2008 R2 cluster. We have zoned 0C ports on Head1 & 2 to top switch and OD ports Head 1 & 2 to bottom switch. We have cluster licenses so running active / active mode.
Today we noticed that when running file operations from head1 that they were being processed through Head 2. THis was reported in Ops Manager. If we looked at FCP Operations / Sec on Head 2 it was up around 2.0K a sec. Head 1 was flat on 0. Looking around people talk about the VTIC internal SCSI bus that is passing the operations from Head 2 to Head1 but i dont know where to go from here.
I thought my switch zones would be ok. I have checked them and confirm all good.
Forgot to note file operations from Head2 are done from Head2 no passing between or offloading the work.
Any help or advice would be great.
2011-10-11 02:42 PM
My first action would be to identify the Controller1 LUNs on the host and check MPIO settings to see which path is preferred. Check out this info on single_image cfmode:
2011-10-12 01:45 PM
Thanks for your help with my VTIC issue and we resolved this yesterday.
Turns out our WWPN (port names) were all duplicated on the 2 x Brocade 500e switches. Once we manually hard set them to be unique we disabled the switch ports re-enabled then bingo DSM mgmt on our file node reported 4 paths and file access was through the correct path. When we downed 0c then 0d paths it would fail to the VTIC and once we up'd 0c and 0d they then became the preferred path again.