Massimo Bollati bio photo

Massimo Bollati

Non e' mai troppo tardi!

Why is IPMI better than snmp?
I received an email asking why I prefer IPMI over snmp.
IPMI is more powerful, more flexible and provides better information… let’s have a look at some examples:

Investigating a restart without logging directly into the ILOM interface:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password chassis restart_cause

System restart cause: chassis power control command  

Checking whether the locate flashing LED is ON or OFF:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem led get LOCATE

LOCATE           | ON  

Event list:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sel list

   1 | 09/28/2015 | 20:35:35 | Power Supply #0xb8 | State Asserted   | Asserted
   2 | 09/28/2015 | 20:35:36 | Power Supply #0xb7 | State Asserted   | Asserted
   3 | 09/28/2015 | 20:35:46 | Power Supply #0xb8 | State Deasserted | Asserted
   4 | 09/28/2015 | 20:35:47 | Power Supply #0xb7 | State Deasserted | Asserted
   5 | 09/28/2015 | 21:17:47 | Power Supply #0xb8 | State Asserted   | Asserted
   6 | 09/28/2015 | 21:17:48 | Power Supply #0xb7 | State Asserted   | Asserted
   7 | 09/28/2015 | 21:17:59 | Power Supply #0xb8 | State Deasserted | Asserted
   8 | 09/28/2015 | 21:18:00 | Power Supply #0xb7 | State Deasserted | Asserted
   9 | 09/28/2015 | 21:39:49 | Power Supply #0xb4 | State Deasserted | Asserted
   a | 09/28/2015 | 21:39:51 | Power Supply #0xb2 | State Deasserted | Asserted
   b | 09/28/2015 | 21:39:51 | System ACPI Power State #0x0f | S0/G0: working | Deasserted

We can check all of them or clear all of them:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sel clear

Clearing SEL.  Please allow a few seconds to erase.

Chaining more commands:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem cli "show /SP/powermgmt/ actual_power" "show /SP/services"

Connected. Use ^D to exit.
-> show /SP/powermgmt/ actual_power
  
/SP/powermgmt
   	 Properties:
actual_power = 741


-> show /SP/services

 /SP/services
    Targets:
http
https
ipmi
kvms
servicetag
snmp
ssh
sso

Check if a PSU has faulted:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem cli "show /SYS/PSU_FAULT"

Connected. Use ^D to exit.
-> show /SYS/PSU_FAULT

/SYS/PSU_FAULT
Targets:

Properties:
type = Indicator
ipmi_name = SYS/PSFAIL/LED
value = On

Check an HW category:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem sdr type "Power Supply"

PS0/VINOK        | B1h | ok  | 10.0 | State Asserted
PS0/PWROK        | B2h | ok  | 10.0 | State Asserted
PS1/VINOK        | B4h | ok  | 10.1 | State Asserted
PS1/PWROK        | B5h | ok  | 10.1 | State Asserted
PS2/VINOK        | B7h | ok  | 10.2 | State Deasserted
PS2/PWROK        | B8h | ok  | 10.2 | State Deasserted
PS3/VINOK        | BAh | ok  | 10.3 | State Deasserted
PS3/PWROK        | BBh | ok  | 10.3 | State Deasserted

List all the categories…:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sdr type

Temperature              | (0x03)  | Fan                       | (0x04)
Physical Security        | (0x05)  | Platform Security         | (0x06)
Processor                | (0x07)  |  Power Supply             | (0x08)
Power Unit               | (0x09)  |  Cooling Device           | (0x0a)
Other                    | (0x0b)  | Memory                    | (0x0c)
Drive Slot / Bay         | (0x0d)  | POST Memory Resize        | (0x0e)
System Firmwares         | (0x0f)  |  Event Logging Disabled   | (0x10)
Watchdog1                | (0x11)  |  System Event             | (0x12)
Critical Interrupt       | (0x13)  | Button                    | (0x14)
Module / Board           | (0x15)  | Microcontroller           | (0x16)
Add-in Card              | (0x17)  | Chassis                   | (0x18)
Chip Set                 | (0x19)  | Other FRU                 | (0x1a)
Cable / Interconnect     | (0x1b)  | Terminator                | (0x1c)
System Boot Initiated    | (0x1d)  | Boot Error                | (0x1e)
OS Boot                  | (0x1f)  | OS Critical Stop          | (0x20)
Slot / Connector         | (0x21)  | System ACPI Power State   | (0x22)
Watchdog2                | (0x23)  |  Platform Alert           | (0x24)
Entity Presence          | (0x25)  |  Monitor ASIC             | (0x26)
LAN                      | (0x27)  |  Management Subsys Health | (0x28)
Battery                  | (0x29)  |  Session Audit            | (0x2a)
Version Change           | (0x2b)  |  FRU State                | (0x2c)

… and choose one:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sdr type 0x0d

DBP/HDD0/FAIL    | 21h | ok  |  4.0 | Predictive Failure Deasserted
DBP/HDD1/FAIL    | 22h | ok  |  4.1 | Predictive Failure Deasserted
DBP/HDD2/FAIL    | 23h | ok  |  4.2 | Predictive Failure Deasserted
DBP/HDD3/FAIL    | 24h | ok  |  4.3 | Predictive Failure Deasserted

Disk Backplane Present:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem cli "show /SYS/DBP/PRSNT"

Connected. Use ^D to exit.
-> show /SYS/DBP/PRSNT
/SYS/DBP/PRSNT
Targets:

Properties:
type = Entity Presence
ipmi_name = DBP/PRSNT
class = Discrete Sensor
value = Present
alarm_status = cleared

Check if the hard Disk0 has failed:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem cli "show /SYS/DBP/HDD0/FAIL"

Connected. Use ^D to exit.
-> show /SYS/DBP/HDD0/FAIL

   /SYS/DBP/HDD0/FAIL
   Targets:

Properties:
type = Drive Slot
ipmi_name = DBP/HDD0/FAIL
class = Discrete Sensor
value = Predictive Failure Deasserted
alarm_status = cleared

We can also filter the queries:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem led get all |egrep -v "OFF|na|OK.*ON"

SYS/POWER/LED    | ON
SYS/ALERT/LED    | ON
SYS/PSFAIL/LED   | ON

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sunoem cli "show /SP/logs/event/list Severity==(critical) -level all -output table"

Connected. Use ^D to exit.
-> show /SP/logs/event/list Severity==(critical) -level all -output table
Target              | Property               | Value                           
--------------------+------------------------+---------------------------------

ID     Date/Time                 Class     Type      Severity
-----  ------------------------  --------  --------  --------
1726   Fri Oct  2 14:19:46 2015  IPMI      Log       critical ID =    4 : 10/02/2015 : 14:19:46 : System ACPI Power State : ACPI : S5/G2: soft-off : Asserted
1725   Fri Oct  2 14:19:46 2015  IPMI      Log       critical ID =    3 : 10/02/2015 : 14:19:46 : System ACPI Power State : ACPI : S0/G0: working : Deasserted
1724   Fri Oct  2 14:19:45 2015  IPMI      Log       critical ID =    2 : 10/02/2015 : 14:19:45 : Power Supply : PS0/PWROK : State Deasserted
1723   Fri Oct  2 14:19:42 2015  IPMI      Log       critical ID =    1 : 10/02/2015 : 14:19:42 : Power Supply : PS1/PWROK : State Deasserted

Check all temperatures:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password sdr type temperature

MB/T_AMB0        | 01h | ok  |  7.0 | 35 degrees C
MB/T_AMB1        | 02h | ok  |  7.0 | 35 degrees C
MB/T_AMB2        | 03h | ok  |  7.0 | 38 degrees C
P0/T_AMB         | 32h | ok  |  3.0 | 20 degrees C
P0/T_CORE        | 33h | ok  |  3.0 | 27 degrees C
P1/T_AMB         | 42h | ok  |  3.0 | 23 degrees C
P1/T_CORE        | 43h | ok  |  3.0 | 25 degrees C
P2/T_AMB         | 52h | ok  |  3.0 | 22 degrees C
P2/T_CORE        | 53h | ok  |  3.0 | 26 degrees C
P3/T_AMB         | 62h | ok  |  3.0 | 21 degrees C
P3/T_CORE        | 63h | ok  |  3.0 | 23 degrees C
P4/T_AMB         | 72h | ok  |  3.0 | 19 degrees C
P4/T_CORE        | 73h | ok  |  3.0 | 27 degrees C
P5/T_AMB         | 82h | ok  |  3.0 | 18 degrees C
P5/T_CORE        | 83h | ok  |  3.0 | 22 degrees C
P6/T_AMB         | 92h | ok  |  3.0 | 19 degrees C
P6/T_CORE        | 93h | ok  |  3.0 | 26 degrees C
P7/T_AMB         | A2h | ok  |  3.0 | 20 degrees C
P7/T_CORE        | A3h | ok  |  3.0 | 21 degrees C

We can even power ON/OFF the server:

ipmitool -I lanplus -H 192.169.4.1 -U my_user -P my_password power soft

Chassis Power Control: Soft

Post a comment

All comments are held for moderation; basic HTML formatting accepted.

Name: (required)
E-mail: (required, not published)
Website: (optional)