IBM Power 8 VIOS problem

[Walter Dey]

I have a customer having  problems with a IBM Power 8 VIOS system which is connected with native FC to a N5k; they see the following error message:

VIOS1:
 Oct  9 10:48:22 fscsi4     T FCP_ERR12           Nameserver GID_PN or GPN_ID failed by adapter driver
 Oct  9 10:48:22 fscsi4     T FCP_ERR4            ioctl() send seq failure for port 0xFFFFFC; ENODEV
 Oct  9 10:48:20 fscsi4     T FCP_ERR12           Nameserver GID_PN or GPN_ID failed by adapter driver
 Oct  9 10:48:20 fscsi4     T FCP_ERR4            ioctl() send seq failure for port 0xFFFFFC; ENODEV
 Oct  9 10:48:17 fscsi8     T FCP_ERR12           Nameserver GID_PN or GPN_ID rejected with reason code 00000013
 Oct  9 10:48:17 fscsi8     T FCP_ERR4            ioctl() send seq failure for port 0xFFFFFC; ENODEV
 Oct  9 10:48:17 fscsi4     T FCP_ERR12           Nameserver GID_PN or GPN_ID rejected with reason code 00000013
 Oct  9 10:48:17 fscsi4     T FCP_ERR4            ioctl() send seq failure for port 0xFFFFFC; ENODEV
 Oct  9 10:48:15 fscsi8     T FCP_ERR12           Nameserver GID_PN or GPN_ID rejected with reason code 00000013

There is a IBM document

http://www-01.ibm.com/support/docview.wss?uid=isg1IZ98294

which describes this error in the context of Cisco interop mode; however, we don't have any interop configuration.

In the N5K I see often the following message

16:11:15 %KERN-3-SYSTEM_MSG: [38620129.811115] fc2_nsh_tx_frame: FC2 s_id/d_id/vsan error: sid=0xfffffe,did=0x0,vsan=100,rctl:0xc3,type:0x0,oxid 0x1,rxid:0xffff - kernel

Can someone explain what this message means ?

[Walter Dey]

problem solved ! eg. connect to a MDS which supports class 2 and 3, or change AIX to only send class 3

https://www.ibm.com/developerworks/community/forums/html/topic?id=77777777-0000-0000-0000-000014962067

AIX server that is either Fibre Channel or FCoE attached to an N5K. AIX is the only device that uses Class 2 for its Fibre Channel Control Plane Traffic. (FICON does as well, but the N5K does not support FICON).