FC-HBA 5716 in IBM p5 series 570 (Fabric DS-4700M SAN Switch problem)

Jump to solution
"invalid attachment" in Operational State while changed Port speed from 4 to 2Gb,
it need to run diagnostic routine to fcs0 to resume Operational State to Online,
is there any parameter need to change to make it stable
Labels (1)
0 Kudos
1 Solution

Accepted Solutions
3 Argentum

Re: FC-HBA 5716 in IBM p5 series 570 (Fabric DS-4700M SAN Switch problem)

Jump to solution
I see this issue with every new AIX machine added to the fabrics. Although my SAN is Brocade based, the behavior seems to be parallel. GlenH's suggestion of deleting the fsc0 device and re-running the cfgmgr does work in our environment. In some cases where the port is stubborn, disabling and renabling that nodes port on the SAN switch helps too. This will cause it to log out and relog back in to the fabric. You can make sure that the settings are correct for the HBA though the smitty screens.

View solution in original post

0 Kudos
3 Replies
3 Argentum

Re: FC-HBA 5716 in IBM p5 series 570 (Fabric DS-4700M SAN Switch problem)

Jump to solution
If its a new connection, you could try removing and rediscovering the device...

# rmdev -d -l fcs0 (you may have to do a recursive delete to remove other devices underneath it also)
Then do
# cfgmgr

If its an existing connection, then i'm not sure if you can change the negotiated speed on-line and you might have to reboot.

AIX HBA's normally automatically negotiate to switched fabric and the correct speed.
0 Kudos
3 Argentum

Re: FC-HBA 5716 in IBM p5 series 570 (Fabric DS-4700M SAN Switch problem)

Jump to solution
Christopher,

invalid attachment means that the port is incompatible with the other device at the other end of the cable.
There are a string of codes behind the event code with could reveal the the reason. Can you post them or is the problem solved?
cheers,

ed
0 Kudos
3 Argentum

Re: FC-HBA 5716 in IBM p5 series 570 (Fabric DS-4700M SAN Switch problem)

Jump to solution
I see this issue with every new AIX machine added to the fabrics. Although my SAN is Brocade based, the behavior seems to be parallel. GlenH's suggestion of deleting the fsc0 device and re-running the cfgmgr does work in our environment. In some cases where the port is stubborn, disabling and renabling that nodes port on the SAN switch helps too. This will cause it to log out and relog back in to the fabric. You can make sure that the settings are correct for the HBA though the smitty screens.

View solution in original post

0 Kudos