Issue with CM mac address and traps , output from CMTS

[startx001]

Hi ,

I have some issue with snmp traps on CMTS uBR10K using PRE2

interface Cable3/0
cable enable-trap cmonoff-notification
cable enable-trap cmonoff-interval 60

snmp-server trap-source Loopback0
snmp-server enable traps cable cm-onoff
snmp-server host x.x.x.x public-cmts
snmp-server host x.x.x.x public-cmts

But I saw that I have a problem with mac addresses .
With debug of snmp packets I recive this

Dec 19 14:41:23.327: SNMP: V1 Trap, ent ciscoDocsExtMIB.2, addr x.x.x.x, gentrap 6, spectrap 1
docsIfCmtsCmStatusEntry.2.4475717 = H{&65 <<<<------- Here should be MAC ??
docsIfCmtsCmStatusEntry.3.4475717 = 10.5.5.87
docsIfCmtsCmStatusEntry.4.4475717 = 0
docsIfCmtsCmStatusEntry.5.4475717 = 8113
cdxCmCpeEntry.5.6.200.251.38.182.53.157 = 2885
cdxCmtsCmStatusExtEntry.1.4475717 = 1
Dec 19 14:41:23.339: SNMP: Queuing packet to x.x.x.x
Dec 19 14:41:23.339: SNMP: V1 Trap, ent ciscoDocsExtMIB.2, addr x.x.x.x, gentrap 6, spectrap 1
docsIfCmtsCmStatusEntry.2.4475717 = H{&65
docsIfCmtsCmStatusEntry.3.4475717 = 10.5.5.87
docsIfCmtsCmStatusEntry.4.4475717 = 0
docsIfCmtsCmStatusEntry.5.4475717 = 8113
cdxCmCpeEntry.5.6.200.251.38.182.53.157 = 2885
cdxCmtsCmStatusExtEntry.1.4475717 = 1
Dec 19 14:41:23.379: SNMP: Packet sent via UDP to x.x.x.x
Dec 19 14:41:23.379: SNMP: Queuing packet to x.x.x.x

Like this one

#scm 10.5.5.87
D
MAC Address IP Address I/F MAC Prim RxPwr Timing Num I
State Sid (dBmv) Offset CPE P
c8fb.26b6.359d 10.5.5.87 C5/1/3/U0 online(pt) 2885!-2.00 823 0 N

Inače normalno bi trebalo da bude ovako

Dec 19 14:42:36.836: SNMP: V1 Trap, ent ciscoDocsExtMIB.2, addr x.x.x.x, gentrap 6, spectrap 1
docsIfCmtsCmStatusEntry.2.4478568 = C8 FB 26 F8 D6 28
docsIfCmtsCmStatusEntry.3.4478568 = 10.5.26.57
docsIfCmtsCmStatusEntry.4.4478568 = 0
docsIfCmtsCmStatusEntry.5.4478568 = 0
cdxCmCpeEntry.5.6.200.251.38.248.214.40 = 5736
cdxCmtsCmStatusExtEntry.1.4478568 = 12
Dec 19 14:42:36.848: SNMP: Queuing packet to x.x.x.x

Looks to me that the debugs are simply printing the values in ASCII or EBCDIC instead of hex:
Mac = c8fb.26b6.359d.
c8= È in ASCII - H in EBCDIC
fb= û in ASCII and EBCDIC
26= & in ASCII
b6= ¶ in ASCII - Þ in EBCDIC
35= 5 in ASCII
9d= (blank space) in ASCII

I believe the IOS will not print special characters (or may replace them for some other characters).
The values above are pretty close to what is seen in the debug.

Also i done snmp walk and there is some ok MAC addresses and some not

SNMPv2-SMI::transmission.127.1.3.3.1.2.1641193 = Hex-STRING: 24 76 7D 99 49 F3
SNMPv2-SMI::transmission.127.1.3.3.1.2.1641199 = Hex-STRING: 10 5F 49 36 1E C1
SNMPv2-SMI::transmission.127.1.3.3.1.2.1641201 = STRING: "$v}J{R"
SNMPv2-SMI::transmission.127.1.3.3.1.2.1641243 = Hex-STRING: C8 FB 26 B5 D6 41

As you can see third in a row.
Why some are written as Hex-STRING: and some as STRING: ??
Anyone with same issue ? Some idea to fix this ?

Thanks

[Manuel Rodriguez]

Hi,

As you mentioned, this is an issue on how the snmp client tool is displaying the result of the poll. This is an snmpwalk/snmpget issue rather than an issue with the result from the Cisco router itself.

I'm not such an expert with snmpwalk/snmpget but, I believe if you run the command using the option '-Ox', this will force output in hex format, and you should see the value as expected.

Ultimately, there should be some MIB configuration in the SNMP client you may need to address in order to get the issue completely solved.

Take a look at this thread: https://groups.google.com/forum/#!topic/mailing.unix.net-snmp-users/-5WfweOMsUE. A similar issue was discussed there (just for a different OID but also involving MAC addresses display) and some hints are provided.

Hope this helps.

Regards.

[startx001]

Hi ,

i done next :

Resets  number of modems and follow after off / on trap messages. Configured that  CMTS is sending  messages to server and my computer. Basically we had the same number of messages that are coming to my comuter and server. We went to the CMTS and there connected computer directly to the CMTS to force sending messages to that PC . We had again identical (although in this case arrived almost all the messages). 

What is missing are the messages when  the modem goes offline (which are more importantly) 

For messages when modem comes online, we recive 99% of messages. 

We observed is that for example in the first 10 modems when we  reset the modem , CMTS send traps only for three modems (not consecutive). After the tenth modem , CMTS send all messages (for test we usually reset 20-30 modems). There is not any connection about the type of modem that sends the trap.

Suspicion fell on PRE card but the tests were tried on various PRE cards (2 CMTS with pre2, 2 with PRE4, and one with pre5). IOS were different from the SCF3 - SCH1.

maybe is  there is a limitation with this setup?, and of course if you have any other solution that will give us info on whether  modems go off / online.

Thanks

[startx001]

I found that could be buffer issue . Does anyone know command to increase snmp/trap buffer on CMTS ?

"For CMTS that is not sending every traps you may change trap buffer value in your configuration (I don't remember how)
If you type "show snmp" you should see dropped traps counter not equal 0. By default they put a very small trap buffer and CMTS drops traps if too many have to be sent at the same time."