SPA 122 hangs up after 600 seconds

braymond1963 · ‎02-11-2013

Hi,

I have this ATA unit that quite often, after 600 seconds, hangs up the call while the FXS is still off-hook. I ran into a similar issue with an SPA9XX phone long time ago where the phone was restarting itself regardless of it being on a call or not.

This is the snippet of the debug output of a call that got hung up:

Feb 11 17:52:23 ACK sip: 1234567890@xxx.xxx.xxx.xxx:5060 ...

Feb 11 17:52:23 modemcable111.111-111-111.mc.videotron.ca

Feb 11 17:52:41 modemcable111.111-111-111.mc.videotron.ca SIP_tsInviteClientEventProc(event: 2)

Feb 11 17:52:41 modemcable111.111-111-111.mc.videotron.ca SIP_sessTsEventProc(event: 31)

Feb 11 17:52:55 modemcable111.111-111-173.mc.videotron.ca SIP_tsInviteClientEventProc(event: 2)

Feb 11 17:52:55 modemcable111.111-111-173.mc.videotron.ca SIP_sessTsEventProc(event: 31)

Feb 11 18:02:27 BITELL-ATA msgswitchd[593]: MSGSWD RTCP Reqt len 12 Data 2,1965716,7312,0

Feb 11 18:02:27 BITELL-ATA [17195123.100000] RTCP is running so calling rtcp stop

Feb 11 18:02:27 BITELL-ATA [17195123.100000] chan->kmode is present not null

Feb 11 18:02:27 BITELL-ATA [17195123.104000] ###### RTCP sock_sendmsg return 172

Feb 11 18:02:29 modemcable111.111-111-111.mc.videotron.ca [1]->xxx.xxx.xxx.xxx: 5060(770)

Feb 11 18:02:29 BYE sip: 1234567890@xxx.xxx.xxx.xxx:5060...

The only parameter that has 600 is the Resync one but i have turned off resync as far as I can tell:

<Provision_Enable>Yes</Provision_Enable>

<Resync_On_Reset>Yes</Resync_On_Reset>

<Resync_Random_Delay>2</Resync_Random_Delay>

<Resync_At__HHmm_/>

<Resync_At_Random_Delay>600</Resync_At_Random_Delay>

<Resync_Periodic>0</Resync_Periodic>

<Resync_Error_Retry_Delay>3600</Resync_Error_Retry_Delay>

<Forced_Resync_Delay>14400</Forced_Resync_Delay>

<Resync_From_SIP>Yes</Resync_From_SIP>

<Resync_After_Upgrade_Attempt>Yes</Resync_After_Upgrade_Attempt>

<Resync_Trigger_1/>

<Resync_Trigger_2/>

<Resync_Fails_On_FNF>No</Resync_Fails_On_FNF>

I now have put Resync_Random_Delay and Resync-At_Random_Delay to a value of 0 to see if the call will hold longer.

I also turned off STUN and kept NAT keep-alive as below:

<Handle_VIA_received>No</Handle_VIA_received>

<Handle_VIA_rport>No</Handle_VIA_rport>

<Insert_VIA_received>No</Insert_VIA_received>

<Insert_VIA_rport>No</Insert_VIA_rport>

<Substitute_VIA_Addr>No</Substitute_VIA_Addr>

<Send_Resp_To_Src_Port>No</Send_Resp_To_Src_Port>

<STUN_Enable>No</STUN_Enable>

<STUN_Test_Enable>No</STUN_Test_Enable>

<STUN_Server/>

<EXT_IP/>

<EXT_RTP_Port_Min/>

<NAT_Keep_Alive_Intvl>60</NAT_Keep_Alive_Intvl>

<Redirect_Keep_Alive>No</Redirect_Keep_Alive>

I wonder if the Substitute_VIA_Addr would have a play here.

What do you think Cisco Support? A bug?

I run 1.3.1(003) and this ATA is behind NAT connected to a D-Link DIR615.

Thanks for your help!

Patrick Born · ‎02-12-2013

Hi braymond1963,

It's certainly not normal for an ATA to reboot at timed intervals. Here are some suggestions in no particular order:

1. You say, "What do you think Cisco Support?" Please note that this site is not "Cisco Support", it's a community of folk trying to help each other out. You can reach Cisco Support here:

http://www.cisco.com/en/US/support/tsd_cisco_small_business_support_center_contacts.html

2. Is there a specific reason that you're changing all of the above-mentioned parameters? The SPA products ship with sane-defaults in that they are factory-set for optimal performance. They don't normally need to be changed from the factory setting. I recommend you factory reset your ATA and start from a known configuration and then only changing the registration parameters to get your ATA to register.

3. Do you have control of your DHCP server? I'm wondering if the ATA's IP address is being changed periodically instead of being renewed, thus forcing it to reboot.

4. You don't mention how the ATA receives its configuration. Are you configuring it by hand using the web-UI or is it being automatically provisioned via DHCP OPTIONS and the ATA's profile rule? If you're using automatic provisioning, set to no the provisioning parameter as follows: No

This will help determine if there's a parameter in the configuration that's causing the reboots.

5. I'm guessing that the reboot problem that plagued your SPA9xx phone could be similar to what's causing your ATA to reboot.

6. Make sure that you're using the 2A power supply that came with your ATA and not a 1A supply from some other product.

Regards,

Patrick---

Use this reference document to locate SPA ATA resources

braymond1963 · ‎02-12-2013

Hi Patrick,

Yes, I should have wrote "Cisco Community" despite the fact I knew Cisco Support do look at the forum; my bad. I also mentioned the reboot thing based on past experience but in this case, I cannot confirm if the ATA actually reboots but I do know it drops the call.

As for the default configs, I am only changing the line specific ones (User ID, Display Name, Password and SIP Port, Register, Proxy). I also need to change the NAT ones as it defaults to a non-NATting implementation so NAT Mapping and KeepAlive are usually turned ON (Yes).

I do think NAT KeepAlive Interval is too low (15 seconds). I can understand that doing it often ensure to keep the hole opened in the router but I think 60 seconds is generally adequate (unless you can prove me otherwise).

This particular ATA hasn`t change IP address so DHCP is not the issue.

The config is fetched during the initial installation from a provisioning server. We pre-provision the units before installation and once the unit boots up, it comes and gets its config. I don't want it to resync unless we send it a NOTIFY so turning Provision_Enable to No is not an option as it prevents a remote resync using SIP. That's the main reason why I chose to set the various resync parameters to 0.

We do use the PSU that comes with the ATA.

So as you can see, it is a fairly basic insall behind a router with the ability to force a resync from SIP.

Note that I am also not excluding an issue with his router. We replaced his first ATA as it was taking 10 seconds before the ATA would start ringing (I created another blog for that last week) and changing to a new ATA fixed it but now this issue ...

Anyway, if you have a recommended config especially around the resync parameters, let me know.

Thanks,

Benoit

braymond1963 · ‎02-13-2013

So looking at the D-Link logs, it drops packets and the call terminates after 10 minutes ... very inconvenient.

I therefore opened UDP ports (506x and 32769-32867) to prevent blocking. But that didn't help on our first test call because the router blocked a packet meant for UDP 5120 ...

Looking at the SIP trace, the calling SPA sets its RTP within the 32769-32867 range but the receiving SPA request it to be 5120!!!

SIP/2.0 200 OK

To: <1234567890>;tag=a74ccd2525ff2e5ai1

From: "Support" <>9876543210@xx.xx.xx.xx>;tag=XXe0tBDpN67Xa

Call-ID: 88456388-f088-1230-9caa-00096bb5998e

CSeq: 40045943 INVITE

Via: SIP/2.0/UDP xx.xx.xx.xx;branch=z9hG4bK5mpUDy137jr2H

Contact: "Test" <1234567890>

Server: Cisco/SPA122-1.3.1(003)

Remote-Party-ID: "Test" <>1234567890@mydomain.com>;screen=yes;party=called

Content-Length: 260

Allow: ACK, BYE, CANCEL, INFO, INVITE, NOTIFY, OPTIONS, REFER

Supported: replaces

Content-Type: application/sdp

v=0

o=- 14380732 14380732 IN IP4 111.111.11.111

s=-

c=IN IP4 111.111.111.111

t=0 0

m=audio 5120 RTP/AVP 0 100 101

a=rtpmap:0 PCMU/8000

a=rtpmap:100 NSE/8000

a=fmtp:100 192-193

a=rtpmap:101 telephone-event/8000

a=fmtp:101 0-15

a=ptime:20

a=sendrecv

So what range of ports does the receiving SPA use? Why is it not using the RTP range defined in config?

Patrick Born · ‎02-13-2013

It's hard to guess what's going on without seeing the ATA's full config and a full, unfiltered packet capture from registration through BYE.

I've tested in my lab and the ATA is behaving well in that:

it uses as the source SIP port, the UDP SIP Port defined in Voice tab > Line N > SIP Settings > SIP Port:

it uses an RTP port in the range defined in Voice tab > SIP > RTP Parameters > RTP Port Min: / RTP Port Max:

To answer your specific question of "So what range of ports does the receiving SPA use?"
In my testing, my ATA uses an UDP port in the RTP range defined at:

Voice tab > SIP > RTP Parameters > RTP Port Min: / RTP Port Max:

Could it be that the SDP that you show is from a device where 5120 is in the allowed range?

Could some NAT be making changes that you're not accounting for resulting in the 5120 port being used? Check by capturing packets on each ATA's WAN port, before traversing any other devices...

Regards,

Patrick---

braymond1963 · ‎02-13-2013

Hi Patrick,

It is a NAT issue from that D-Link DIR615 router. The SIP Debug I get on my syslog server from that ATA do show the audio port to be in range:

v=0

o=- 14380732 14380732 IN IP4 192.168.0.199

s=-

c=IN IP4 192.168.0.199

t=0 0

m=audio 32847 RTP/AVP 0 100 101

...

But my wireshark shows that SIP message (once it traversed NAT) to have not only the IP changed (which is expected) but also the audio port ... (doh)

Thanks for all the tips!

Benoit

Patrick Born · ‎02-13-2013

Hi Benoit,

You're welcome, thanks for letting the Community and me know.

Regards,

Patrick---