SL3000 drives on windows media server recognized as StandAlone and Robotic drives

[hytham_fekry]

We Have the following configuration :

- Node1,Node2 "clustered master servers" where Node2 is active 

- node 2,3,4 configured as media servers on which STK SL3000 with 6 D10000D drives configured , each server  have 2 HBA .

Zones is configured in the same way on the 3 servers :

Zone 1 : HBA1 PWWN , 6 tape drives WWN . "Drives Port A selected from storageTek SLC console"

Zone 2 : HBA 2 PWWN , 6 tape drives WWN . "Drives Port B selected from storageTek SLC console" , Robot WWN .

Node3 is controlling the robot , Node 2,4 are configured to access the robotic unit through remote host  .

When configuring node4 using configuration wizard , the 6 tapes drives on SL3000 is seen twice :

-  Once (On port 5) under robot control "Where SN is the same as seen in SLC console and as seen from node 2,3 , The target no, is the same on all hosts" 

- Once (On port 6 ) as a standalone drive  "strangely , Drives here have different SN from SN seen in SLC console/node2,3  and have different target no. .. that's why NBU mistakenly think it's different drives .

From O.S side I removed ghost devices for robot/drives , restarted device media manager on node 4 and 2 ,  robot is deteced fine on device media manager from windows , with 12 paths for the 6 drives  . 

Please Advise 

[sdo]

Are you able to share the zone config for any zones that use the initiators? Not the entire zoning 'config/cfg', just the zones that have your tape initiators within them. If soft zoning, then are you able to also share the aliases (if used), for all of the members of the selected zones that you share wth us? Pls not all aliases, just the ones pertinent to the zones that contain your tape initiators. Single fabric? Single initiator zoning? Definitely no disk and tape zoned to the same initiators? You use the term 'node2' twice - can you clarify/re-describe the node names?

[hytham_fekry]

I'm just using node1/node2 names for simplicity .. let's say train1 , train2 , etc ..

train1 , train2 nodes are clustered master servers , train2 is active .

train3 , train 4 are media servers ...

The problem as described above is only on train4 .. 

All 4 servers have exactly the same configuration where each server have two dual ported HBA , "customer uses WWN no. directly with no aliases in soft zone configuration" , cutsomer uses one  WWN from  the Robot  , and  2 WWN for each tape drive "mapped to 2 HBAs" :

Zone1 : HBA1/Port1 ===> mapped to SAN storage .

zone2 : HBA1/port2 ,  6 T10000D tape drives WWNs "Port A is selected from each drive"  .

Zone3 : HBA2/Port1 ===> mapped SAN Storage .

zone4 : HBA2/port2 , 6 T10000D tape drives WWNs ."Port B is selected from each drive"  , Robot intiator WWN .

Train3 controls the robot , train2 , train4 are configured to use train3 as the remote host who is controling the robot .

Only Train4 , See 6 tapes twice , once as stand alone (where tapes have different target no. and oddly different SN )

and once under robot control (where tapes have same target no. and SN as train2 and train3)

* Should robot intiator WWN be mapped to train2 , train4 ?

* I removed ghost devices for robot and drives  on  train4 , will remove it from train3 next visit ..

Any Ideas ?