No outgoing call after migration on Nexus

[euanetwork]

Hi,

Since we migrated our Call manager on the nexus infrastructure we cannot place any outgoing calls.

The gateways are seen in an unknown state which seems to prevent the CCM from routing outgoing calls.

The incoming calls work fine.

Any idea what could mess up this ?

thx!

[Cedric Van Labeke]

Hi Xavier,

Is there full network connectivity between IP Phones, CUCM and outbound gateways?

What do you mean exactly with Nexus infrastructure? VMWare & Nexus 1000v or... ?

What is meant with outbound calls are failing? They drop upon connect? You can't form the complete number even? There's no call drop but dead-air?

Thanks,

Cedric

[euanetwork]

Hello Cedric

- I'm Stéphane, a colleague from Xavier -

We use here two CCM (one sub, one pub), one Unity  and one  Voice router as a gateway. Two E1 cards are used separately in this voice router, one for the CCM outside calls (in & out), the other for a Fax system (not going through the CCM) : as the Fax system required H323, we have not implemented MGCCP, that's why the gateway is  displayed in "unknown mode"  on the CCM web interface.

The CCMs, Unity and Voice Gateway are in the same VLAN and  in the same IP block  as described below :

Voice VLAN 102  - IP block 10.64.28.0/24

CCM1 : 10.64.28.1

CCM2 : 10.64.28.2

Unity : 10.64.28.3

Voice Gateway : 10.64.28.5

IP Phones (for ex 7940G)  :  10.64.28.10 to 10.64.28.200 (DHCP)

Gateway (on the Nexus 7010) : 10.64.28.254

When we start to move all the devices (CCMs, Unity, Voice Gateway and IP phones on POE Switches) are all plugged on the N7K (even the same card) we suddenly loose the possibility to place external calls (fast dial tone "busy" ring) while the incoming calls function still work perfectly.

The ip connectivy remain perfectly fine during all this migration (except of course the unplug/replug of the servers) and after the migration : each device can ping the others ones.

To answer your questions :

Is there full network connectivity between IP Phones, CUCM and outbound gateways? YES

What do you mean exactly with Nexus infrastructure? VMWare & Nexus 1000v or... ?  only ONE 7010 in this case

What  is meant with outbound calls are failing?

     They drop upon connect?   Fast dial tone busy

     You  can't form the complete number even? No we can complete the number

     There's no call drop but dead-air?   Fast dial tone busy is the answer I think

=> The only workaroung we have found to retrieve the outside calls function is the cold reboot of the voice gateway (the 2851 router)

Let me know what is missing for your understanding, I will update you

Thanks

Stéphane

[Cedric Van Labeke]

Hi Xavier,

Nice to meet you. So if I read all your comments below I think we can conclude this:

- Migration of devices of some previous switch to new Nexus 7010, even same card.

- All devices are in same VLAN and there's full IP connectivity

- Inbound calls to your IOS Voice Gateway still work.

- You can form the entire number but then you hear fast busy

- You are using MGCP to control the E1 for inbound and outbound calls (correct? or also H323?)

- There's another E1 in the Gateway for a fax solution which you use H323 for

- It appears that a reboot of the Gateway resolves this.

New questions:

1. How long after a reboot does it continue to work fine the outbound calls?

2. Have you ever tried to make more than 30 calls after one failing call rather than rebooting the gateway?

3. How long does it take until you have formed the number and you get the fast busy signal?

4. What do you see in the "debug isdn q931" for a bad call compared to a good call?

5. When you migrated these devices to the new switch. Have you had to poweroff other devices? (think of unsaved configs)

Cheers,

Cedric

[euanetwork]

Cedric, iIt's Stéphane again this time, not Xavier

One point to correct  : we are ONLY using H323 on the gateway (so H323 for the CCM and H323 for the Fax system) ; as it is a "global" parameter (placed on the LAN  interface of the Gateway), we have no other choice I believe

interface GigabitEthernet0/0
  description * ACCESS-VLAN-102-ToIP *
  ip address 10.64.28.5 255.255.255.0
  h323-gateway voip interface
  h323-gateway voip h323-id NANTES02-1
  h323-gateway voip bind srcaddr 10.64.28.5

I was just taking about MGCP as this was the protocol we would have used IF the Fax systems would have support it (which was not the case)

My answers to the new questions:

1. How long after a reboot does it continue to work fine the outbound calls? it works as long as we do no "touch" anything among the voice equipments. For example, if we just unplug/replug the Unity server we are almost sure that the problem will immediately reappear.  (we know that the unity server IS not concerned by the process of sending / receiving a call but I can really  reproduce the pb doing this...)

=> This means that we currently have  all theses devices plugged on the 7010 and working good ...just waiting for the next problem.

2. Have you ever tried to make more than 30 calls after one failing call rather than rebooting the gateway? No, approx 10 calls max.  You would like to have 30 concurrents calls or just 30 attempts ?

3. How long does it take until you have formed the number and you get the fast busy signal?  "immediately"  up to  one/two second

4. What do you see in the "debug isdn q931" for a bad call compared to a good call?  we have not performed this specific debug but theses two ones :

-       debug voip ccapi inout
-       debug cch323 h225

and the result was this one :

when it does not works :

NANTES02-1#
*Mar 14 19:28:50.648: //-1/xxxxxxxxxxxx/H323/cch323_h225_receiver: Received msg of type SETUPIND_CHOSEN
*Mar 14 19:28:50.648: //-1/xxxxxxxxxxxx/H323/setup_ind: Entry
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: callingNumber[272640341] calledNumber[0610271532]
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: ---- calling IE present
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: ====== PI = 0
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: Receive: infoXCap 0
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: Receive: infoXCap ccb 0
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind:
setup_ind: is_overlap = 0, info_complete = 0

*Mar 14 19:28:50.648: //17626/802AA4274502/H323/setup_ind: Call Manager detected
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/cch323_h225_receiver: SETUPIND_CHOSEN: src address = 10.64.28.5; dest address = 10.64.28.2
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/run_h225_sm: Received event H225_EV_SETUP_IND while at state H225_IDLE
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/common_idle_setupInd_hdlr: no match is found
*Mar 14 19:28:50.648: //17626/802AA4274502/H323/cch323_h225_set_new_state: Changing from H225_IDLE state to H225_WAIT_FOR_ARQ state
*Mar 14 19:28:50.652: //17626/802AA4274502/H323/run_h225_sm: Received event H225_EV_RAS_FAILED while at state H225_WAIT_FOR_ARQ
*Mar 14 19:28:50.652: //17626/802AA4274502/H323/cch323_h225_send_release: Cause = 3; Location = 0
*Mar 14 19:28:50.652: //17626/802AA4274502/H323/cch323_h225_send_release: h225TerminateRequest: src address = 171973637; dest address = 10.64.28.2

when it works fine

NANTES02-1#
*Mar 14 21:49:22.287: //-1/xxxxxxxxxxxx/H323/cch323_h225_receiver: Received msg of type SETUPIND_CHOSEN
*Mar 14 21:49:22.287: //-1/xxxxxxxxxxxx/H323/setup_ind: Entry
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: callingNumber[272640341] calledNumber[0610271532]
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: ---- calling IE present
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: ====== PI = 0
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: Receive: infoXCap 0
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: Receive: infoXCap ccb 0
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind:
setup_ind: is_overlap = 0, info_complete = 0

*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/setup_ind: Call Manager detected
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/cch323_h225_receiver: SETUPIND_CHOSEN: src address = 10.64.28.5; dest address = 10.64.28.2
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/run_h225_sm: Received event H225_EV_SETUP_IND while at state H225_IDLE
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/common_idle_setupInd_hdlr: no match is found
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/cch323_h225_set_new_state: Changing from H225_IDLE state to H225_SETUP state
*Mar 14 21:49:22.287: //21/808EBCC4B102/H323/cch323_create_incoming_callinfo_block: peer is NULL - may affect modem pass through! ccb: 449464BC, ccNewCallInfo 44F52320

5. When you migrated these devices to the new switch. Have you had to poweroff other devices? (think of unsaved configs) We have tried to reboot the CCMs, the voice routers and the unity server   but - at the end - only a reboot of the router (voice gateway) fixed the issues after having established the complete ip connectivity.

Last : our gateway is the following one :

System image file is "flash:c2800nm-ipvoice-mz.124-25d.bin"

Cisco 2851 (revision 53.50) with 249856K/12288K bytes of memory.

Regards

Stephane

[Cedric Van Labeke]

Hi Guys

Ok, so only H.323 that's fine. You could mix MGCP and H.323, that's not a problem.

I'm very much intrigued by what you are seeing here... .

You're stating that almost all the time (reproducible even), when you unplug the network cable of your Unity server, you can't make outbound calls anymore. And each and every time it fails you see this in the debug:

*Mar 14 19:28:50.648:  //17626/802AA4274502/H323/cch323_h225_set_new_state: Changing from  H225_IDLE state to H225_WAIT_FOR_ARQ state
*Mar 14 19:28:50.652: //17626/802AA4274502/H323/run_h225_sm: Received event H225_EV_RAS_FAILED while at state H225_WAIT_FOR_ARQ
*Mar 14 19:28:50.652: //17626/802AA4274502/H323/cch323_h225_send_release: Cause = 3; Location = 0
*Mar  14 19:28:50.652: //17626/802AA4274502/H323/cch323_h225_send_release:  h225TerminateRequest: src address = 171973637; dest address = 10.64.28.2

So you're outbound call from CallManager, goes to a Gatekeeper? Can you clarify the flow? As I understood it like:

IP Phone -- (SIP/SCCP) -- CUCM -- (H.323) -- GW -- (E1 PRI ISDN) -- PSTN

Where does the gatekeeper fit in?

If it's expected to be there, what do you see on your gatekeeper traces?

I can't see the link between unplugging a cable from an unrelated server and RAS ARQ failing.

Is your network speed and duplex correct on both ends of the wire for all devices?

[euanetwork]

- Stephane again -

The unplugging/replugging of Unity server was just an example : the problem can be reproduce doing the same with CCM1, CCM2 or gateway  (that's why I was using the term of "touching" ie moving - bad french translation -  any device on this voice vlan)

Now your questions :

So you're outbound call from CallManager, goes to a Gatekeeper?  => No

Can you clarify the flow? As I understood it like:

IP Phone -- (SIP/SCCP) -- CUCM -- (H.323) -- GW -- (E1 PRI ISDN) -- PSTN

=> it's correct at it is , I would just replace CUCM by CCM but it's the same

Where does the gatekeeper fit in?    No gatekeeper

If it's expected to be there, what do you see on your gatekeeper traces?   N/A

I can't see the link between unplugging a cable from an unrelated server and RAS ARQ failing.

=> same ... but this is really  the only thing done

Is your network speed and duplex correct on both ends of the wire for all devices? We have checked this part too, it's all correctly configured.

Regards

Stéphane

[Cedric Van Labeke]

If there's no gatekeeper, why do we then see RAS in the H323 statemachine debugs for H225?

And if there's no gatekeeper, why do you have an h323-id configured on your Network interface:

-> h323-gateway voip h323-id NANTES02-1