Cisco UCS vic 1227 single point of failure

[bnk0000011]

Hi,

One quick question. Is the Cisco UCS vic 1227 a single point of failure. I ask because we have just updated our Cisco server to 3.1.2b, and it failed badly. It looks like the MLOM update has failed, and now we cannot make it to work again, so the server is stuck in an bade state. The server has only one MLOM, so now we have no communication between the server and Fabric Interconnect. MLOM have one connection toFI A and one connection to FI B.

Is it recommended to have 2 MLOM installed, in case of an update fails on one MLOM, then we still have the other?

Is the 2 MLOM ports two separate ports, or is there a single point of failure behind them?

[Walter Dey]

- Was this setup configured according to

http://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/c-series_integration/ucsm2-2/b_C-Series-Integration_UCSM2-2/b_C-Series-Integration_UCSM2-2_chapter_0110.html#reference_EF9772524CF3442EBA65813C2140EBE6 ?

- Are the versions compliant with above ?

- can you please give more details for the failing update ? eg. did you do a firmware autoinstall and the server autoinstall ?

[bnk0000011]

Yes and Yes;)

UCS and Fabric Interconnect was done by auto install, and the server was updatet using Host Firmware Package. No mater what i try the FSM stops at Setting of NIV mode failed. I have tried the following on the server: Re-acknowledge, Decommission, Reset MIMC/CMOS. Removed server from Service Profile, and added it to the service profile again. Downgradet nicke, bios, but, still the same FSM message:(

But do you know it the MLOM is a single point of failure?

[Walter Dey]

Do I understand you correctly ?

- Autoinstall infrastructure went ok ?

- which UCS version ?

- Autoinstall server failed ? was this happening for all your servers ? blade and rack, or only rack ?

- please post the error message ?

- which FI and IOM model ?

And yes, some components of VIC-1227 are single point of failures.

[bnk0000011]

Yes. auto install of UCS and Interconnect ... no problems.

They are now on ver. 3.1.2b.

We tried to upgrade 4 servers, to 3.1.2b also, and all went wrong. We have only rack mounted servers UCSC-C240-M4SX.

FI model UCSC-C240-M4SX

MLOM is vic 1227 model.

Error message: Setting of NIV mode failed (Error code 1002). It tries to set the NIV mode about 15-20 times, and then it starts all over Again, and ends with the error above.  It goes like that forever.

[Walter Dey]

Are the servers directly connected to the fabric interconnect with the 10G ports ? or are you using fex's ?

can you connect with UCS Manager to the CIMC ?

Still don't understand this error ? is it seen with UCSM ? and FSM is stuck ? and if yes, in which step (updating ??)

[bnk0000011]

Yes, I can connect with UCS Manager to CIMM.

The error is seen in UCS Manager under the tab FSM. It is not stuck, but tries Step 55 many times, and then starts from step 1 Again. And does that repeatedly.

[Walter Dey]

I assume this situation happens, if you apply the new Host firmware package ?

Which component update creates this error ? BIOS, CIMC, I/O adaptor,.....

What do you see in the error log of UCSM ?

[bnk0000011]

Yes, the error started after we applyed new Host firmware package 3.1.2b.

We do not know which component that generates this erro.

The SEL log is not logging anything after the update.

It is repporting1 Critical errors

1. [FSM:FAILED]: rack-Mount server discovery 1 (FSM: sam:dme:ComputerRackUnitDiscover). Remote-Invocation-Error: FSM Retries Exhausted. ID 618084.

[Walter Dey]

I think you tried

decommission / recommission ?

btw.

from which version did you do the upgrade ?

You still have servers which are working ok ?