What is the speed of the ports on your upstream switch ? We have B200 M3 blades and we get the same 20 GB .I leave it auto with no issues . We have our uplinks port channeled. ... View more

I would make sure all the paths look good going to your FC switches . You can check the logs on Brocades if your using them or MDS switches. Make sure you have primary and secondary targets in your boot policy . Check your hosts do they have Dead paths on your Fiber connections ? When ESxi loads up ... it loads into memory . So even a disconnect might not be painful but if its trying to write to a debug log on the LUN that might cause a crash cause it cannot get there . Just my 2 cents . Chris ... View more

Update : Our 2 FI's only had one PSU plugged in and according to the Tech-support file  FI-A is pointing to a power lose. I since plugged in both ports each FI making sure that they are on separate circuits . When the FI seemed to crash it didn't produce a core file . We will know in a couple of days if its the issue . TAC has been really helpful ... View more

Something interesting ... On Saturday the primary FI crashed and I left it alone as the subordinate . It has not crashed since . I am going to call TAC this morning though but that is interesting . ... View more

Hi Abhinav , I did the upgrade last Thursday night at 11:00 . The first crash was on Friday at 2:40 pm . The second crash was Saturday at 4:43 pm . After the first and second time the primary changes to the secondary like its supposed to do after a crash . The one that is crashing is the primary but after the last vrash I left it the secondary now . This is the error code : fltNetworkElementInoperable Fault Code:F0291 Message Fabric Interconnect [id] operability: [operability] Explanation This fault typically occurs when the fabric interconnect cluster  controller reports that the membership state of the fabric interconnect  is down, indicating that the fabric interconnect is inoperable. Recommended Action If you see this fault, take the following actions: Step 1 Verify that both fabric interconnects in the cluster are running the same Kernel and System software versions. Step 2 Verify that the fabric interconnects software version and the Cisco UCS Manager software versions are the same. Step 3 If the above actions did not resolve the issue, create a show tech-support file and contact Cisco TAC. When the fiber interconnect shuts off all the ports shutdown . I wonder when it crashes does it do a gARP so that the second FI takes over ? I'm not sure what I can do ? Thanks Chris ... View more