Deconfigured Cpu'S

[mjos]

On my rx8640 server, when I do a parstatus it shows that two CPU's are deconfigured.
How do I configure them.

[CPU Details]

Type : N/A
Speed : 1418 MHz
CPU Status
=== ======
0 OK
1 OK
2 Deconfigured
3 Deconfigured
4 OK
5 OK
6 OK
7 OK

CPUs
===========
OK : 6
Deconf : 2
Max : 8

Any info will he highly helpfull - thanks

[Jeeshan]

The reasons for a deconfigured CPU could be various like i/d cache h/w failures and/or exceeding the LPMC thresholds of i/d cache errors, bus problem etc. These problems would make the CPU as "not responding", the firmware will detect the "degradation" and it will de-configure it. This all happens during CPU self test routines on power-on as well while system is up and running OS.

is there HP Integrity Management Agent installed there?

how many cells in there?

these 2 processors are in same cell?

[mjos]

There are 2 active cells

CPU Memory Use
OK/ (GB) Core On
Hardware Actual Deconf/ OK/ Cell Next Par
Location Usage Max Deconf Connected To Capable Boot Num
========== ============ ======= ========= =================== ======= ==== ===
cab0,cell0 Active Core 8/0/8 32.0/0.0 cab0,bay0,chassis0 yes yes 0
cab0,cell1 Active Base 6/2/8 32.0/0.0 cab0,bay0,chassis1 yes yes 0
cab0,cell2 Absent * - - - - - -
cab0,cell3 Absent * - - - - - -

Notes: * = Cell has no interleaved memory.

There are 14 processors which makes it dual boot

nh1d0t01#ioscan -fnC processor
Class I H/W Path Driver S/W State H/W Type Description
========================================================================
processor 0 0/120 processor CLAIMED PROCESSOR Processor
processor 1 0/121 processor CLAIMED PROCESSOR Processor
processor 2 0/122 processor CLAIMED PROCESSOR Processor
processor 3 0/123 processor CLAIMED PROCESSOR Processor
processor 4 0/124 processor CLAIMED PROCESSOR Processor
processor 5 0/125 processor CLAIMED PROCESSOR Processor
processor 6 0/126 processor CLAIMED PROCESSOR Processor
processor 7 0/127 processor CLAIMED PROCESSOR Processor
processor 8 1/120 processor CLAIMED PROCESSOR Processor
processor 9 1/121 processor CLAIMED PROCESSOR Processor
processor 10 1/124 processor CLAIMED PROCESSOR Processor
processor 11 1/125 processor CLAIMED PROCESSOR Processor
processor 12 1/126 processor CLAIMED PROCESSOR Processor
processor 13 1/127 processor CLAIMED PROCESSOR Processor

How do i check whether these two processors are part of same cell?

[mjos]

Yes both the deconfigured CPU's are part of the same cell - cell 1

nh1d0t01#parstatus -c 1 -V
[Cell]
Hardware Location : cab0,cell1
Global Cell Number : 1
Actual Usage : Active
Normal Usage : Base
Connected To : cab0,bay0,chassis1
Core Cell Capable : yes
Firmware Revision : 7.48
Failure Usage : Normal
Use On Next Boot : yes
Partition Number : 0
Partition Name : Partition 0 - UX
Requested CLM value : 0.0 GB
Allocated CLM value : 0.0 GB
Cell Architecture Type : Itanium(R)-based
CPU Compatibility : CDG-578
Hyperthreading Capable : yes
GSM Capable : yes

[CPU Details]

Type : N/A
Speed : 1418 MHz
CPU Status
=== ======
0 OK
1 OK
2 Deconfigured
3 Deconfigured
4 OK
5 OK
6 OK
7 OK

[Jeeshan]

all right then. check the firmware of that cell board. may be some mismatch there.

BTW, if your OS claimed all the CPU's, check with the 'top' command.

[mjos]

TOP is showing 14 CPUs

Cpu states:
CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS
0 0.44 6.8% 0.0% 1.5% 91.7% 0.0% 0.0% 0.0% 0.0%
1 0.40 9.3% 0.0% 3.1% 87.6% 0.0% 0.0% 0.0% 0.0%
2 0.51 38.1% 0.0% 12.4% 49.5% 0.0% 0.0% 0.0% 0.0%
3 0.45 7.7% 0.0% 3.5% 88.8% 0.0% 0.0% 0.0% 0.0%
4 0.44 34.6% 0.0% 7.7% 57.6% 0.0% 0.0% 0.0% 0.0%
5 0.40 17.6% 0.0% 4.8% 77.6% 0.0% 0.0% 0.0% 0.0%
6 0.46 26.9% 0.0% 2.7% 70.4% 0.0% 0.0% 0.0% 0.0%
7 0.48 25.7% 0.0% 15.7% 58.6% 0.0% 0.0% 0.0% 0.0%
16 0.61 14.5% 0.0% 3.3% 82.2% 0.0% 0.0% 0.0% 0.0%
17 0.57 35.2% 0.0% 24.0% 40.8% 0.0% 0.0% 0.0% 0.0%
18 0.53 23.2% 0.0% 1.7% 75.0% 0.0% 0.0% 0.0% 0.0%
19 0.49 7.0% 0.0% 0.8% 92.3% 0.0% 0.0% 0.0% 0.0%
20 0.47 23.0% 0.0% 16.6% 60.3% 0.0% 0.0% 0.0% 0.0%
21 0.82 57.3% 0.0% 31.3% 11.4% 0.0% 0.0% 0.0% 0.0%
--- ---- ----- ----- ----- ----- ----- ----- ----- -----
avg 0.50 23.4% 0.0% 9.3% 67.3% 0.0% 0.0% 0.0% 0.0%

Cell 0 has 8 CPUs & Cell 1 has 8 CPUs.
Since two CPUs in Cell 0 is showing as deconfigured, TOP is shoiwng as 14 CPUs

[mjos]

How do I check the firmware of the cell board. Is there any way...I can check or modify the status or the CPU from the EFI shell

[sujit kumar singh · Accepted solution]

Hi

1)the Firmware for all the parts of the SD can be checked logging onto the MP Console.

GO To the COmmand Menu.

MP>CM
from there give command sysrev.


MP:CM>sysrev

this shall showe all trhe firmwares of the SD that is PDC,PDHC,CELL,IO etc.

2)the other part is that the Deconfigured CPUs can be made to use on the next boot only.

3) as this is an SD, Firmware (upgrade if reqd) and the CPU operations better to get done by logging a Case with HP and Authorosed HP Engineers only.


Regards
Sujit

[likid0]

Check on the FPL logs on the mp, and search for this type of message:

CELL_HW_DEGRADED


if you see these type messages, you have a HDW issue on the cpus

[likid0]

Sorry i didn't see your other post, you can try and enable the cpus from the efi:


cpuconfig

to see if the CPUs are shown as de-configured.

If this is what has been done, then:

cpuconfig 0 on
cpuconfig 1 on
cpuconfig 2 on

[sujit kumar singh]

hi


just in add of what orange_adm says the
#cpuconfig on

considering this an SD

might take the form as

EFI Shell>cpuconfig on

to activate the Deconfigured processors.

where CELL_NO can take the form of 0/1 , 1

etc that means 0/1 is in the Cell local format and reads Cabinet0/Cell1 and later reads local format that is Simply cell 1.

Obviously for a multicabinet SD.


Regards
Sujit


Regards
sujit

[rick jones]

In addition to being deconfigured as part of predictive, and being explcitily deconfigured via cpuconfig, isn't it possible that the cpus are iCOD? I don't know what that would look like at the various levels but thought I would mention it.

[mjos]

Thanks, I was able to make the CPU's up through the EFI menu.
However is there any way to check whether these CPU's are part of iCod

[sujit kumar singh]

Hi,


to check if the CPUs on the system are a part of ICOD or ICAP, the following commands can be tried.
#icapstatus
#icod_stat


#icapstatus

Regards
Sujit

[mjos]

Below is the icod_stat output.
Does it specify in the below output...

#icod_stat
Software version: B.08.01
System ID: TX1APDS1
Serial number: ########
Product number: AQ90NW
System contact e-mail: Not set
From e-mail: Set to the default ('adm')
Asset reporting: on
Temporary capacity warning period: 15 days
Exception status: No exception

Local nPartition Status
-----------------------
Total number of configured cores: 16
Number of Intended Active cores: 16
Number of active cores: 16
Number of inactive cores: 0
Additional cores that can be activated with current usage rights: 0
Number of cores that could be activated with additional usage rights: 0
Number of cores that can be activated with temporary capacity: 0
Number of cores that are deconfigured or attached to inactive cells: 0

Instant Capacity Resource Summary
---------------------------------
Number of cells without usage rights: 0
Number of inactive cells: 0
Amount of memory without usage rights: 0.0 GB
Amount of inactive memory: 0.0 GB
Number of cores without usage rights: 0
Number of inactive cores: 0
Number of cores that must be deactivated (insufficient usage rights): 0
Temporary capacity available: 0 days, 0 hours, 0 minutes

Allocation of Instant Capacity Resources among the nPartitions
--------------------------------------------------------------
Intended Actual
nPar Total Active Active =======Inactive======= Runs
ID Cores Cores Cores Cores Memory Cells iCAP nPar Name
==== ===== ======== ====== ====================== ==== ======================
0 16 16 16 0 0.0 GB 0 Yes Partition 0... (local)
N/A 0 N/A N/A N/A 0.0 GB 0 N/A Unassigned Cells

[sujit kumar singh]

Hi

can you also immediately post the icod_stat -v

Regards
Sujit

[sujit kumar singh]

Hi


and
also

#icod_stat -u

these shall be the ICOD logs.
Regards
Sujit

[mjos]

Hi,

There is no icod_stat -u or icod_stat -v

#icod_stat -u
Usage: icod_stat
icod_stat -s