IO Backplane Link Down

[Cy Hauptmann]

Hi all,

I was wondering on how do I go about getting my link on IO Module backplane.

My set up as follows ..

7 B200 server and 1 B440 server in a chassis.  All of my backplane ports are online except for 1 of the B440 ports.  It just has a link down.  Now it's not the same on both fabrics.  On Fab A it's port 1/8, on Fab B it port 1/7.  I have no error messages to go along with this link-down.

I have tried to reset the DCE interfaces on the server but that hasn't work.  I also have resetted the IO Module and that hasn't work.

What am I missing??

Thanks, in advance.

Cy.                  

[Robert Burns]

The IOM "backplane" links come up automatically when there's a service profile configured.  These aren't something you have control enabling/disabling.

Second, you can't fit 7x B200's (1/2 width server) and a B440 (full width server) in a single Chassis.  There's a total of eight 1/2 width slots in each chassis.  A full width server such as the B440 would take up 2 x 1/2 width slots.

The B440 has two Mezz slots.  So unless you have both Mezz cards populated all the IOM backplane ports for this server will not be up.

Please confirm how many adapters you have in the B440.  Equipment - Chassis - Servers - ServerX - Adaptors.

Regards,

Robert

[Cy Hauptmann]

Sorry, meant to say 6 b200s .. .and yes my b440 has the two mezz cards in them.  I have two adapters in my b440.

I have two chassis with b440s on it them .... on the second chassis, only 1 port down on the fabric, fabric a .. fabric b is good to go .....

With no errors to go by, I'm at a loss on what to do next.

[padramas]

Hello Cy,

Has this port ever worked ?

Please share UCSM version, FI ,IOM and adapter model

To start with provide the output of

show service-profile circuit server x/y

==> where  X = chassis and Y = blade slot

connect nxos a

show int eth X/1/

Padma

[jonathan.liu]

Hi Padma,

I apologize for hijacking this thread Cy.

On behalf of our customer one of the big movie houses, We are encountering the same issue with one of our backplane ports on FI-A being down. On one of our blade servers (server 8) Adapter fc port 1 link went down mysteriously on January 23^rd of this year. Before which everything was up and UCS reported no errors.

As far as the setup goes: We have a UCS Mini that has four B200M4 blades, each FI has 2 ethernet ports configured as a port channel connected to a Juniper core switch and 2 ports configured as standalone uplinks to Brocade SAN.

To answer your questions:

1. Yes the port was up before January 23, 2017.
2. UCSM: 3.1(2e)
3. 6324 FI: ucs-mini-k9-system.5.0.3.N2.3.12e.bin
4. IOM: 4.1(2d)
5. Adapter: Cisco UCS VIC 1340

# show service-profile circuit server 1/8

Service Profile: xxxxxx

Server: 1/8

    Fabric ID: A

        Path ID: 1

        VIF        vNIC            Link State  Oper State Prot State    Prot Role   Admin Pin  Oper Pin   Transport

        ---------- --------------- ----------- ---------- ------------- ----------- ---------- ---------- ---------

               811 eth0            Unknown     Unknown    No Protection Primary     0/0/0      0/0/0      Ether

               814 eth1            Unknown     Unknown    No Protection Backup      0/0/0      0/0/0      Ether

               815 fc0             Admin Down  Admin Down No Protection Unprotected 0/0/0      0/0/0      Fc

              9007                 Unknown     Unknown    No Protection Unprotected 0/0/0      0/0/0      Ether

    Fabric ID: B

        Path ID: 1

        VIF        vNIC            Link State  Oper State Prot State    Prot Role   Admin Pin  Oper Pin   Transport

        ---------- --------------- ----------- ---------- ------------- ----------- ---------- ---------- ---------

               812 eth0            Up          Active     Active        Backup      0/0/0      0/0/1      Ether

               813 eth1            Up          Active     Active        Primary     0/0/0      0/0/1      Ether

               816 fc1             Up          Active     No Protection Unprotected 0/0/0      1/0/2      Fc

              9008                 Up          Active     No Protection Unprotected 0/0/0      0/0/0      Ether

Its peculiar because when physically checking the switch, all LED lights are green, from UCSM, it shows the port as operable and up. But there is a major fault error: F0207 "Adapter fc host interface 1/8/1/1 link state: down.” On the SAN side, it shows the port as up. On UCS side, int vfc 815 is showing as err-disabled.

To troubleshoot:

• Re-acknowledged the server and adapter.
• I physically re-seated the blade and re-acknowledged the slot.
• I reseated the SFP and fiber cable.
• I reset adapter 1 connectivity.
• Disabled and re-enabled the port.
• I moved a VM running W2K12 onto the server.

Customer has purchased SMARTNET but it hasn’t been activated yet due to delay which is out of our control. Any insight you can provide would be most appreciated.

Thank you and Happy New Year!

Jonathan