Sun Microsystems, Inc.  Oracle System Handbook - ISO 7.0 May 2018 Internal/Partner Edition
   Home | Current Systems | Former STK Products | EOL Systems | Components | General Info | Search | Feedback

Asset ID: 1-72-1583034.1
Update Date:2017-08-01
Keywords:

Solution Type  Problem Resolution Sure

Solution  1583034.1 :   Oracle Fabric Interconnecct (Formerly Xsigo) PVI (Private Virtual Interface) vnic Shows 'up/failed' Status on One Fabric Interconnect  


Related Items
  • Oracle Fabric Manager
  •  
  • Oracle Fabric Interconnect F1-15
  •  
  • Oracle Fabric Interconnect F1-4
  •  
Related Categories
  • PLA-Support>Sun Systems>SAND>Network>SN-SND: Oracle Virtual Networking
  •  




In this Document
Symptoms
Cause
Solution


Applies to:

Oracle Fabric Interconnect F1-15 - Version All Versions and later
Oracle Fabric Interconnect F1-4 - Version All Versions and later
Oracle Fabric Manager - Version 4.0.0 to 4.0.2 [Release 4.0]
Information in this document applies to any platform.

Symptoms

When creating HA (High Avaiability) PVI vnics, when the HA PVI vnic pair is deployed to two interconnected (meshed) Fabric Interconnects creating one large meshed IB Fabric, one of the PVI HA vnic pair shows state up/failed on one Fabric Interconnect.  This is observed when using Oracle Fabric Manager below version 4.1.1.   (OFM versions 4.1.0 and below).

Cause

This is caused by the IO Profile (server-profile from Fabric Interconnect CLI) being terminated to the same HCA port GUID on both Fabric Interconnects instead of IO Profile on Interconnect A being terminated to one HCA port GUID and the other IO Profile getting terminated to the second HCA Port GUID. 

Solution

 From both Fabric Interconnect CLI interfaces when logged in as user 'admin' run the 'show physical-server' command:

show physical-server

frack.us.oracle.com 24e89097ff452a ExtSw-2c902004126e8-Port17 VMware/ESXi-5.3.0.ESX.1/x86_64 2.7.5559/3.0.0 frack

frack.us.oracle.com 24e89097ffa4d4 ExtSw-2c90200419858-Port17 VMware/ESXi-5.3.0.ESX.1/x86_64 2.7.710/3.0.0

In the above example CLI output, note that server-profile 'frack' is terminated to HCA port GUID ending in '452a' on Fabric Interconnect 'arkansas'.  

In this example, the second Fabric Interconnect is 'texas' and looking at server-profile (IO Profile) 'frack' this is seen:

show physical-server

frack.us.oracle.com 24e89097ffa4d4 ExtSw-2c90200419858-Port17 VMware/ESXi-5.3.0.ESX.1/x86_64 2.7.710/3.0.0 frack

frack.us.oracle.com 24e89097ff452a ExtSw-2c902004126e8-Port17 VMware/ESXi-5.3.0.ESX.1/x86_64 2.7.5559/3.0.0

When using Oracle Fabric Interconnect versions below 4.1.1, it is possible that both the 'frack' server-profiles (IO Profile is OFM) end up terminated to the same HCA port GUID.  For instance both 'frack' server-profiles end up terminated to to HCA port GUID ending in 452a instead of frack server-profile on Fabric Interconnect 'arkansas' being terminated to HCA Port GUID ending in 452a and 'frack' server-profile on Fabric Interconnect 'texas' being terminated to HCA Port GUID 'a4d4'. 

To resolve this issue, perform this workaround:

From either Fabric Interconnect perform:

set server-profile disconnect <server-profilename>

Answer yes to the warning, then reconnect the server-profile to the unused HCA port GUID:

set server-profile <profilename> connect 

In the above command after typing 'connect' hit double tab to bring up server-profile list, then copy and paste the line that matches the unused HCA Port GUID then hit enter.  This will then connect the server-profile to the previously unused HCA port GUID and the HA PVI vnic will then show state as up/up.

This is a known resolved bug in Oracle Fabric Manager versions below 4.1.1.

 

 

 

 


Attachments
This solution has no attachment
  Copyright © 2018 Oracle, Inc.  All rights reserved.
 Feedback