C3750X - Power supply OIDs problems

Markelson · ‎10-09-2012

Hello,

We are trying to monitor two stacks of CISCO WS-c3750X running IOS 12.2.(58)SE.

We can get the state of the power supplies, but after a power cycle or reboot of the switches the identifier of ONE of the Power supplies ALWAYS change. The OIDs for the rest of the PS do not vary. The problem happens in the two stacks.

Any idea or work around?

Kind regards

Mark

deyadav · ‎10-31-2012

Hi Mark,

Few other users are also experiencing this problem and I am currently working on one of the TAC case around the same. I'll update you how it goes, however at this moment I'm trying to establish if this is expected or an unexpected behavior. If possible, try not to poll individual OID indexes for the power supply and instead use the OID only, example:

server% snmpwalk -v1 -c public 1.3.6.1.4.1.9.9.13.1.5.1.2

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1034 = STRING: "Sw1, PS1 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1035 = STRING: "Sw1, PS2 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.2034 = STRING: "Sw2, PS1 Normal, RPS NotExist"

I'll update once there is more insight on this.

Regards,

Deepak

Lasse Nylander · ‎02-05-2013

Hi any update on this issue?

Just updated software, problem not solved yet

Switch Ports Model SW Version SW Image

------ ----- ----- ---------- ----------

* 1 30 WS-C3750X-24 15.0(2)SE1 C3750E-UNIVERSALK9-M

2 30 WS-C3750X-24 15.0(2)SE1 C3750E-UNIVERSALK9-M

-

snmpwalk -v 2c -c ********** X.X.X.X .1.3.6.1.4.1.9.9.13.1.5.1.2

iso.3.6.1.4.1.9.9.13.1.5.1.2.1034 = STRING: "Sw1, PS1 Normal, RPS NotExist"

iso.3.6.1.4.1.9.9.13.1.5.1.2.1036 = STRING: "Sw1, PS2 Faulty, RPS NotExist"

iso.3.6.1.4.1.9.9.13.1.5.1.2.2034 = STRING: "Sw2, PS1 Normal, RPS NotExist"

iso.3.6.1.4.1.9.9.13.1.5.1.2.2035 = STRING: "Sw2, PS2 Faulty, RPS NotExist"

-

#sh env al

..

SW PID Serial# Status Sys Pwr PoE Pwr Watts

--- ------------------ ---------- --------------- ------- ------- -----

1A C3KX-PWR-350WAC LIT161700PF OK Good Good 350/0

1B C3KX-PWR-350WAC AZS161800V4 OK Good Good 350/0

2A C3KX-PWR-350WAC AZS161800QS OK Good Good 350/0

2B C3KX-PWR-350WAC AZS161800V9 OK Good Good 350/0

..

vincent.morton · ‎04-24-2013

Deepak,

Do you have an update on this?

We have the same issue and the monitoring we use needs to poll individual OID values in order to get a numerical response, the

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1034 = STRING: "Sw1, PS1 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1035 = STRING: "Sw1, PS2 Normal, RPS NotExist"

type output is no use to us for monitoring purposes.

Thanks

deyadav · ‎05-12-2013

Apologies for delay in responding. Here is the final update although may not solve purpose easily for most NMS systems, however the following is the status:

IOS is unable to maintain a persistent state for the following OID indexes:

entPhysicalIndex

This is used to provide index for power supplies as well which may change after reload. So, the requirement for NMS is to ensure discovering the correct index after reload and then use for monitoring power supplies.

Unfortunately, this change has to happen on NMS and IOS cannot accomodate this. AFAIK IOS-XR got some persistency through enhancement bug:CSCsl67774, however with IOS is seen as a limitation due to design.

Thanks & Regards,

Deepak

asafayan · ‎05-30-2013

Hi,

I'm trying to determine if the issue addressed in this post is a match with the following symptom:

As you can see from the output below - the switch 2 power supplies appear to be healthy:

SWITCH#show env power switch 1

SW PID Serial# Status Sys Pwr PoE Pwr Watts

-- ------------------ ---------- --------------- ------- ------- -----

1A C3KX-PWR-350WAC AZS161707F3 OK Good Good 350/0

1B C3KX-PWR-715WAC LIT16180RRU OK Good Good 715/0

SWITCH#show env power switch 2

SW PID Serial# Status Sys Pwr PoE Pwr Watts

-- ------------------ ---------- --------------- ------- ------- -----

2A C3KX-PWR-715WAC LIT16180RRY OK Good Good 715/0

2B C3KX-PWR-350WAC AZS161707FZ OK Good Good 350/0

Yet, the snmpwalk output shows a Sw2 PS2 power supply fault:

snmpwalk -v1 X.X.X.X -c READONLY .1.3.6.1.4.1.9.9.13.1.5.1

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1035 = STRING: "Sw1, PS1 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.1036 = STRING: "Sw1, PS2 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.2034 = STRING: "Sw2, PS1 Normal, RPS NotExist"

SNMPv2-SMI::enterprises.9.9.13.1.5.1.2.2035 = STRING: "Sw2, PS2 Faulty, RPS NotExist"

Do my symptoms match with the issue addressed in this post ?

TIA,

Amir

Pacerfan9_2 · ‎06-18-2013

I am not understanding how a NMS would detect a change, is that some functionality that can be built into the software? Was anyone able to come up with a solution or workaround to monitor power supply status after reboot?

a.olsthoorn · ‎08-13-2013

Hi Deepak,

can you tell us what solution is suggested by TAC suppport?

Is this a know bug?

We see this on Prime infra 1.3 as well.

kind regards

Ad

deyadav · ‎08-14-2013

This is a limitation imposed by IOS and no solution from IOS perspective to overcome this. Upon reload the indexes would not be persistent and rediscovery from NMS is required to find the right object/index and then monitor it, if this could be performed by NMS. I am not a NMS expert, so feel free to post your question on a new forum to seek help to make this possible from NMS applications.