Graphing DSP channel usage

[chryzo]

I'm stuck trying to graph DSP channel usage. Other VoIP templates like "VoIP Channels in Use on Cisco AS5300 and like Access Routers" counts the serial / isdn channels in use, not the DSP channels. What I am after is graphing of something similar to the "show voice dsp active" - or to be more specific; the SNMP OID's cdspInUseChannels and cdspActiveChannels. If I something useful out of those OID's, you could compare them with active E1-channels and active calls, to see if you got either E1-channel hang or DSP channel hang.

My problem is that the DSP OID's reports numbers that I cannot understand or make sense out of, ex:
On cacti server:
user@server~$ snmpwalk -v1 -c public 10.10.10.10 .1.3.6.1.4.1.9.9.86.1.2.1.1.8
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.10 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.11 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.25 = Gauge32: 21513
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.26 = Gauge32: 4294953011
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.27 = Gauge32: 4294961518
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.28 = Gauge32: 4294965938

While the router shows:
VG1#sh voice call st | inc active
11 active calls found
VG1#sh voice dsp active

DSP DSP DSPWARE CURR BOOT PAK TX/RX
TYPE NUM CH CODEC VERSION STATE STATE RST AI VOICEPORT TS ABORT PACK COUNT
==== === == ======== ======= ===== ======= === == ========= == ===== ============

----------------------------FLEX VOICE CARD 0 ------------------------------
*DSP ACTIVE VOICE CHANNELS*
DSP DSPWARE VOX DSP SIG DSP PAK TX/RX
TYPE VERSION CODEC NUM CH TS VOICEPORT SLT NUM CH TS RST AI ABRT PACK COUNT
===== ======= ======== === == == ========= === === == == === == ==== ============
C5510 4.4.708 g711ulaw 001 01 31 0/1/0:15 000 002 14 31 0 0 0 17187/17480
C5510 4.4.708 14400 001 02 30 0/1/0:15 000 002 13 30 0 0 0 1356/16301
C5510 4.4.708 g711ulaw 001 03 29 0/1/0:15 000 002 12 29 0 0 0 256/282
C5510 4.4.708 g711ulaw 001 04 28 0/1/0:15 000 002 11 28 0 0 0 5403/5526
C5510 4.4.708 g711ulaw 001 05 27 0/1/0:15 000 002 10 27 0 0 0 46451/47219
C5510 4.4.708 g711ulaw 001 06 26 0/1/0:15 000 002 09 26 0 0 0 715/747
C5510 4.4.708 g711ulaw 002 01 25 0/1/0:15 000 002 08 25 0 0 0 1371/1419
C5510 4.4.708 g711ulaw 002 02 24 0/1/0:15 000 002 07 24 0 0 0 641/682
C5510 4.4.708 g711ulaw 002 05 21 0/1/0:15 000 002 04 21 0 0 0 9733/9919
C5510 4.4.708 g711ulaw 002 06 20 0/1/0:15 000 002 03 20 0 0 0 3903/3998
C5510 4.4.708 g711ulaw 003 05 14 0/1/0:15 000 001 14 14 0 0 0 7959/8119
------------------------END OF FLEX VOICE CARD 0 ----------------------------


This was from a gateway with 96 DSP channels (6 x 16). Anyone got any easy way to keep track of the DSP channels?

[chryzo]

Testing against another gateway with ipvoicek9-mz.124-15.T4 (the hardware is more or less the same), I get correct values:
VG1#sh voice dsp active

DSP DSP DSPWARE CURR BOOT PAK TX/RX
TYPE NUM CH CODEC VERSION STATE STATE RST AI VOICEPORT TS ABORT PACK COUNT
==== === == ======== ========== ===== ======= === == ========= == ===== ============


----------------------------FLEX VOICE CARD 0 ------------------------------
*DSP ACTIVE VOICE CHANNELS*
DSP DSPWARE VOX DSP SIG DSP PAK TX/RX
TYPE VERSION CODEC NUM CH TS VOICEPORT SLT NUM CH TS RST AI ABRT PACK COUNT
===== ========== ======== === == == ========= === === == == === == ==== ============
C5510 9.4.4 g711ulaw 001 01 01 0/0/0:15 000 001 01 01 0 0 0 4120/3770
C5510 9.4.4 g711ulaw 001 02 02 0/0/0:15 000 001 02 02 0 0 0 890/784
C5510 9.4.4 g711ulaw 001 03 03 0/0/0:15 000 001 03 03 0 0 0 7235/6918
C5510 9.4.4 g711ulaw 001 04 04 0/0/0:15 000 001 04 04 0 0 0 11215/11519
C5510 9.4.4 g711ulaw 001 05 07 0/0/0:15 000 001 07 07 0 0 0 11125/11029
C5510 9.4.4 g711ulaw 001 06 05 0/0/0:15 000 001 05 05 0 0 0 28578/28577
C5510 9.4.4 g711ulaw 002 01 06 0/0/0:15 000 001 06 06 0 0 0 3452/3257
C5510 9.4.4 g711ulaw 002 02 08 0/0/0:15 000 001 08 08 0 0 0 2787/2883
C5510 9.4.4 g711ulaw 002 03 09 0/0/0:15 000 001 09 09 0 0 0 2644/2219
C5510 9.4.4 g711ulaw 002 04 10 0/0/0:15 000 001 10 10 0 0 0 2583/1589
C5510 9.4.4 g711ulaw 002 05 11 0/0/0:15 000 001 11 11 0 0 0 2567/2363
C5510 9.4.4 g711ulaw 007 01 12 0/0/0:15 000 001 12 12 0 0 0 20564/20784
C5510 9.4.4 g711ulaw 007 02 13 0/0/0:15 000 001 13 13 0 0 0 155/39
C5510 9.4.4 g711ulaw 008 01 19 0/0/0:15 000 002 02 19 0 0 0 9874/9971
C5510 9.4.4 g711ulaw 008 02 20 0/0/0:15 000 002 03 20 0 0 0 9787/9572
C5510 9.4.4 g711ulaw 008 03 21 0/0/0:15 000 002 04 21 0 0 0 9246/8925
------------------------END OF FLEX VOICE CARD 0 ----------------------------


server:~$ snmpwalk -v 2c -c public 10.10.10.10 .1.3.6.1.4.1.9.9.86.1.2.1.1.8
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.14 = Gauge32: 6
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.15 = Gauge32: 5
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.16 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.17 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.20 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.21 = Gauge32: 0
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.22 = Gauge32: 2
SNMPv2-SMI::enterprises.9.9.86.1.2.1.1.8.23 = Gauge32: 3

[chryzo]

Hmkai. Finally I managed to graph my precious DSP channels. Anyone interested in templates, give me a PM.

[mcutting]

Definately interested - would you mind posting them ?

Tried to PM you, but for some unknown reason, it's been stuck in the outbox for ages..

[chryzo]

I finally got a reply from Cisco TAC (and bumping myself yet again ):
There seems to be a bug in IOS prior to 124-5 that causes SNMP to report wrong data. I haven't tested this myself by upgrading one of the gateways I'm having trouble with, but I'll get back to you as soon as I have confirmed that spservices is behaving as nicely as ipvoice.
On the topic of my graphs - I'm having trouble figuring out a nice way to summarize all my DSP resources in one graph - if anyone have any tips please feel free to modify.

[chryzo]

The bugID is CSCeg55420 - not very informative, but still a bugID

[skinty]

[chryzo]

This is more like what I wanted to get - information on all DSP channels in one graph. To bad I had to make a nasty shell-script . . . :/