snmp issue

Good evening list

I have an snmp issue on an ASR, the NMS reports it as uncontactable..

it seems to be generating packets to udp 161 but no replies are seen from
the WAN interface and the ones that are seen are coming in on an ibgp
connection with the other CE,

debug snmp pack

*Feb 12 01:37:28.937: SNMP: Packet sent via UDP to x.x.x.x
*Feb 12 01:37:28.942: SNMP: Packet sent via UDP to x.x.x.x

debug snmp detail

Not sure if I should be worried about below??

SrV2GenerateNotification:Function has reached clean up routine.
*convert_pdu:Conversion between V2 and V1 notification failed.*
CheckMIBView: OID is in MIB view.
CheckMIBView: OID is in MIB view.
CheckMIBView: OID is in MIB view.
CheckMIBView: OID is in MIB view.
CheckMIBView: OID is in MIB view.
SrCheckNotificationFilter: OID is included.
SrCheckNotificationFilter: OID is included.
SrCheckNotificationFilter: OID is included.
SrCheckNotificationFilter: OID is included.
SrCheckNotificationFilter: OID is included.
Sr_send_trap: trap sent to x.x.x.x:162

now the device can reach both NMS servers via icmp ok via the source
loopback0 of the traps & the configs have been checked 100 times to be same
as the other working CE partner so it's not this... also the config has
been re-applied to eliminate

hostname#show snmp
Chassis: (omitted)
1252189 SNMP packets input
    0 Bad SNMP version errors
    10326 Unknown community name
    7 Illegal operation for community name supplied
    0 Encoding errors
    11322135 Number of requested variables
    0 Number of altered variables
    991671 Get-request PDUs
    246181 Get-next PDUs
    7 Set-request PDUs
    0 Input queue packet drops (Maximum queue size 1000)
2413792 SNMP packets output
    0 Too big errors (Maximum packet size 1500)
    9894 No such name errors
    0 Bad values errors
    0 General errors
    1238167 Response PDUs
    1171929 Trap PDUs

SNMP logging: enabled
    Logging to x.x.x.x.162, 0/10, 578622 sent, 6496 dropped.
    Logging to x.x.x.x.162, 0/10, 578622 sent, 6496 dropped.

the dropped counters, unknown community name & no such name errors is
historic and has not incremented since t-shooting....& input/outputs
increment ok

also it seems to queue the udp it send to the NMS where as the "working" CE
does not

*Feb 12 12:07:41.157: SNMP: Queuing packet to x.x.x.x

when it receives an snmp udp it receives them on the CE-CE iBGP link

*Feb 12 12:09:26.667: SNMP: Packet received via UDP from x.x.x.x on
GigabitEthernet0/0/7.1

and for some reason starts generating a V1 trap

*Feb 12 12:08:51.142: SNMP: V1 Trap, ent ciscoMgmt.187, addr loopback0,
gentrap 6, spectrap 1

now having checked the trace to the NMS it goes over the eBGP WAN link &
both NMS servers ip's are learnt via eBGP

it's almost like the loopback0 source of the snmp traps on the CE is not
seen from the WAN side

don't have jurisdiction that far..

also don't have access to the NMS to check from that side but is their
anything else that I can gauge your suggestions on..

thanks in advance

Tony

Blogs and organic groups at http://www.ccie.net
Received on Tue Feb 12 2013 - 19:27:35 ART