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: 1-72-1534629.1
Update Date:2014-11-18
Keywords:

Solution Type  Problem Resolution Sure

Solution  1534629.1 :   Sun Blade X6275 node denied power up due to "Exceeds blade max power of 510 watts" error.  


Related Items
  • Sun Blade 6048 System
  •  
  • Sun Blade 6000 System
  •  
  • Sun Blade X6275 Server Module
  •  
Related Categories
  • PLA-Support>Sun Systems>x86>Blades>SN-x64: BLADE
  •  


Sun Blade X6275 node denied power up due to "Exceeds blade max power of 510 watts" error.

In this Document
Symptoms
Changes
Cause
Solution


Applies to:

Sun Blade 6000 System - Version All Versions and later
Sun Blade 6048 System - Version All Versions and later
Sun Blade X6275 Server Module - Version All Versions and later
Information in this document applies to any platform.

Symptoms

We have seen at numerous sites where after a chassis power outage, the blades power back on and are denied permission to power on from the CMM due to an incorrect Max Power limit being set:

This can be seen from the example messages below:

6462 Mon Feb 11 08:54:31 2013 Power Log minor
/CH/BL4 allocated 255 watts
6461 Mon Feb 11 08:54:28 2013 Power Log major
/CH/BL4 denied request for 570 watts: Exceeds blade max power of 510 watts
6460 Mon Feb 11 08:29:58 2013 Power Log minor
/CH/BL4 allocated 255 watts
6459 Mon Feb 11 08:29:55 2013 Power Log major
/CH/BL4 denied request for 570 watts: Exceeds blade max power of 510 watts
6458 Mon Feb 11 08:29:53 2013 Power Log major
/CH/BL4 denied request for 570 watts: Exceeds blade max power of 510 watts
6457 Mon Feb 11 08:29:29 2013 Power Log minor
/CH/BL4 allocated 255 watts
6456 Mon Feb 11 08:29:26 2013 Power Log major
/CH/BL4 denied request for 570 watts: Exceeds blade max power of 510 watts

Changes

 N/A

Cause

The X6275 Server is a two node blade and each node has to be configured exactly the same (symmetrically).  Because of this, on initial power up, only node 0 communicates with the CMM and presents what power is needed for both nodes which sets the max power limit automatically.  During initialization of the blade node 0, ILOM checks it's configuration over the I2C bus on the motherboard and then doubles that power usage amount to account for node 1 as well.  In some cases, node 0 doesn't initialize completely before the ILOM probes the I2C bus to get the configuration information after an AC power cycle of the chassis.  Because the system isn't initialized all the way, it can sometimes miss important components in the power estimate needed to run, usually CPUs, but can be DIMMs too. 


When this happens, ILOM on node 0 requests an amount of power to run that is to low for both nodes to run.  What ends up happening is the nodes attempt power up and of course all components get configured into the system and the ILOM of that node again requests what the real power is that is needed for that specific node, in this case 315 watts.  But because the max amount was set to the incorrect lower amount 510w, it is below what it will actually take to run both sides.  When the second node powers up, it does not have enough power allotment available from the CMM and so is denied power up permission.  This can be seen in the logs above.  The max power that was requested is 510 watts but the blade actually needs 630 watts to run both nodes.  So the first node powered up and took 255 of the 510 allotment which only leaves 255.  The second node comes up and requests the correct 315 and is denied because there is only 255 left and the 255+315=570 is more than was originally allotted.

Solution

 Workaround:

To work around this issue reset the blade SP.:

-> cd /SP

-> reset

 
Then reset the CMM so it takes affect:

-> reset /CMM

 
This is happening due to a bug in the CMM ILOM which is not allowing the blade SP to update it's power budget on the PCA9501 eeprom.  This will be fixed in a future release of CMM firmware SW3.3.x.  

If after the SP reset the blade still isn't allowed to power on the customer should try reseating the blade.  If the blade still doesn't power up then something else is wrong, possibly an faulty hardware component.  An SR should be opened to investigate further via Services.

The bug associated with this issue is #16424352.


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