Hi, We have UCS blades which are cabeled and i can see on fabric with green light. Server team has connected ESX hosts to that UCS ports. I have zoned them to VMAX array. When i list logins in VMAX array i am seeing that the host is not logged in. Can you help me how to trouble shoot this. Below is the output from the list logins.
Director Identification : FA-14E
Director Port : 0
User-generated Logged On
Identifier Type Node Name Port Name FCID In Fabric
---------------- ----- --------------------------------- ------ ------ ------
20000025b511b18f Fibre XXXesx2209 HBA2 9d0103 No Yes
Solved! Go to Solution.
Make sure you have enabled NPIV on the switch side. If it's Cisco MDS switch run show feature and you can find out.
Also make sure that ports that are connected from FI to MDS switch are in the right VSAN. Do you see a flogi for VMAX FA (14-E0) on the fabric? Make sure that the port it's connected on the MDS switch is on the right VSAN. Shut and noshut the ports to make sure.
Hope this helps.
be careful doing "shut" on the switch, if NPIV is working there are multiple FLOGI connecting through the port ...you shut that port and they all lose connectivity.
Some OS'es / HBA drivers will log out of FC targets if there is no IO activity or no active volumes on those targets... Can you see the targets/LUNs from the host side (e.g. with the inq utility)?
Sent from my iPhone
Have you tried allocating some storage from VMAX and try to write it to that device just try some 2 TB allocated if I/O is established you can see the logins.
You can make sure that the zoning by checking that is on the correct VSAN, on both the fabrics.
Check the switch port for FlogI and whether you can see the Right WWNs.
I didn't catch the mention of ESX (reading comprehension fail on my part)... but the fact that there's an entry in the FA login table at all means that the host WWN did login to that FA at one point -- so either the zoning is correct, or it was correct at one point and something has been changed/broken, or ESX has logged itself out of the VMAX FA target due to lack of IO activity. So in addition to basic zoning/FC path troubleshooting, I'd check things from a host perspective as well -- try to discover VMAX devs down that path with the vSphere Client, and/or generate some IO from the host side and see if it logs back in.