Serious network trouble sip xmit returns error for pkt data

Здравствуйте, на сервере установлен дистрибутив Sangoma Linux с FreePBX 14.0.13.4 периодически отваливаются SIP подключения у менеджеров, перезапуск сервиса астериска не помогает. Все начинает работать только после перезагрузки сервера.
IP адресов менеджеров нету в черном списке iptables.
Связь у менеджеров идет через webcrt ( библиотека sipml5 на фронте )
Из-за чего может падать подключение?

Вот часть логов астериска на момент сбоя ( менеджеры пожаловались на то что отвалилась связь в промежутке времени с 10:20 до 10:25 )

[2019-07-25 09:40:07] ERROR[16477] iostream.c: Problem setting up ssl connection: error:00000001:lib(0):func(0):reason(1), Internal SSL error

[2019-07-25 10:21:53] ERROR[10051] chan_sip.c: Serious Network Trouble; __sip_xmit returns error for pkt data

Если сеть не работает , то и перезапуск сервера не поможет.

Здравствуйте, XOLLER!
Я периодически жалуюсь всем подряд на
1) боли в области поясницы (правда!)
2) невнимательность участников форума, которые приходят с проблемами, логами, вопросами, ответы на которые обычно лежат в этих самых логах (как у вас сейчас)
3) инертность пользователей, заказчиков, которым надо-надо прямо вчера ещё, а потом пропадают в туман и тишину.

Посоветуйте по этим трём проблемам что-то?

ded писал(а):Я периодически жалуюсь всем подряд на

и я тоже жалуюсь, а мне помогут ?
Особенно на пользователей/заказчиков, которые смотрят в лог, где разработчик старался, выводил ошибку, а пользователь/заказчик в розовых очках видимо.

XOLLER писал(а):[2019-07-25 10:21:53] ERROR[10051] chan_sip.c: Serious Network Trouble; __sip_xmit returns error for pkt data

какое из 3х выделенных слов вам перевести ?
То что вы подключаетесь к серверу по SSH это конечно здорово, но не показатель того, что с сетью все в порядке. Вам же лог черным по белому пишет о том, что есть проблема.
Поэтому снимите розовые очки, смотрите в логи и не только астериска, но и самой системы, а так же лог свича и статистику порта куда воткнут сервер.
Открою вам ещё одно заморское слово — ТРАБЛШУТ (troubleshoot).

res_http_websocket.c: WebSocket connection from ‘**.***.**.***:44722’ for protocol ‘sip’ accepted using version ’13’

Для WebRTC нужно многое. Например понимание что это такое и как оно работает…

I’m having these 3 issues after a fresh iciBox_v8_1.x86_64-8.1.2 install:

1) Asterisk is not starting on boot. If I start manually it works.

2)I keep getting this on CLI:

sip_reliable_xmit: Serious Network Trouble; __sip_xmit returns error for pkt data

3)No sound when logging in using WebRTC viciphone.

Here’s the CLI:

[Oct 31 10:42:44] == WebSocket connection from ‘12.226.124.15:56590’ for protocol ‘sip’ accepted using version ’13’
[Oct 31 10:42:45] ERROR[4442]: tcptls.c:447 tcptls_stream_close: SSL_shutdown() failed: error:00000005:lib(0):func(0):DH lib, Underlying BIO error: Broken pipe
[Oct 31 10:42:45] == WebSocket connection from ‘12.226.124.15:56307’ forcefully closed due to fatal write error
[Oct 31 10:42:45] — Registered SIP ‘2525’ at 12.226.124.15:56590

I do not hear the initial «You are the only……»

Issue seems to be with the certificate (I’ve used vicibox-certbot). I did not get any errors while installation.

ERROR[4675]: tcptls.c:447 tcptls_stream_close: SSL_shutdown() failed: error:00000005:lib(0):func(0):DH lib, Underlying BIO error: Broken pipe

Any ideas?

I get this immediately after login:

Agent Screen:

No one is in your session: 8600051

WebRTC log shows:

REGISTERED

2018-10-31 13:45:31 =>
displayName: 2525
uri:

2525@66.XXX

.X.146
authorizationUser: 2525
password: m5Y4FLgd543543weeWWW
wsServers:

wss://dialer.certf.com:8089/ws

CLI:

ERROR[4675]: tcptls.c:447 tcptls_stream_close: SSL_shutdown() failed: error:00000005:lib(0):func(0):DH lib, Underlying BIO error: Broken pipe

Did you enter the location of your certificates in /etc/apache2/vhosts.d/1111-default-ssl.conf

SSLCertificateFile /etc/certbot/live/FQDN/cert.pem
SSLCACertificateFile /etc/certbox/live/FQDN/fullchain.pem
SSLCertificateKeyFile /etc/certbox/live/FQDN/privkey.pem

Now I’m getting this:

viciphone debug log:

2018-10-31 15:21:53 =>
displayName: 2525
uri:

2525@66.58.85.146

authorizationUser: 2525
password: m5Y4FLQOUfROo2weeWWW
wsServers:

wss://dialer.crft.com:8089/ws

2018-10-31 15:21:56 => Got Invite from <0000000000> «ACagcW15410136786666666666666666»
2018-10-31 15:21:56 => Auto-Answered Call

Call gets auto-answered on viciphone. But here’s the CLI:

[Oct 31 15:12:26] == WebSocket connection from ‘12.333.333.65:49729’ for protocol ‘sip’ accepted using version ’13’
[Oct 31 15:12:26] — Registered SIP ‘2525’ at 12.XXX.XXX.65:49729
[Oct 31 15:12:26] NOTICE[3405]: chan_sip.c:24639 handle_response_peerpoke: Peer ‘2525’ is now Reachable. (105ms / 2000ms)
[Oct 31 15:12:29] == Manager ‘sendcron’ logged on from 127.0.0.1
[Oct 31 15:12:29] == DTLS ECDH initialized (automatic), faster PFS enabled
[Oct 31 15:12:29] == Using SIP RTP CoS mark 5
[Oct 31 15:12:29] — Called 2525
[Oct 31 15:12:29] — SIP/2525-00000003 is ringing

asterisk is not detecting the answer.

[Oct 31 20:43:08] == WebSocket connection from ‘12.226.XX.65:57378’ for protocol ‘sip’ accepted using version ’13’
[Oct 31 20:43:08] — Registered SIP ‘2525’ at 12.226.XX.65:57378
[Oct 31 20:43:08] > Saved useragent «VICIphone 1.0-rc1» for peer 2525
[Oct 31 20:43:08] NOTICE[22371]: chan_sip.c:24639 handle_response_peerpoke: Peer ‘2525’ is now Reachable. (100ms / 2000ms)
[Oct 31 20:43:11] == Manager ‘sendcron’ logged on from 127.0.0.1
[Oct 31 20:43:11] == DTLS ECDH initialized (automatic), faster PFS enabled
[Oct 31 20:43:11] == Using SIP RTP CoS mark 5
[Oct 31 20:43:11] — Called 2525
[Oct 31 20:43:11] — SIP/2525-00000000 is ringing

WebRTC log:

2018-10-31 20:43:44 =>
displayName: 2525
uri:

2525@66.XXX

.4.146
authorizationUser: 2525
password: m5Y4FLQOUfRygyg666gg
wsServers:

wss://dialer.clrt.com:8089/ws

2018-10-31 20:43:47 => Got Invite from <0000000000> «ACagcW15410329896666666666666666»
2018-10-31 20:43:52 => Answered Call

if vicibox-certbot ran without any real issues then it should have set the correct SSL certificates in /etc/asterisk/http.conf and /etc/apache2/vhosts.d/1111-default-ssl.conf.

The only thing you really need to do is edit the ‘WebRTC’ template to point to the certificates as well. So I would double check that and see if that’s your issue. You can find it under Admin —> Templates. It looks like you already got the WSS set correctly since it’s registering.

Other then that I’m not sure. Some of the issues sound network related. All 3 issues can be attributed to network issues but that doesn’t mean it’s their cause. There is an init script log at /var/log/vicidial.log that should show you what the init script was seeing during start up. Item 2 and 3 almost 100% describe a NAT traversal issue.

I’m also getting this error on the asterisk console while testing vicibox 8.1.2:

Serious Network Trouble; __sip_xmit returns error for pkt data

It happens after a while when you are logged in as agent with webphone, not immediately.

It seems to be WebRTC related:

https://community.freepbx.org/t/chan-si … a/39534/16

I also found this:

In chan_sip, a “__sip_reliable_xmit: Serious Network Trouble; __sip_xmit returns error” error typically means that there’s a misconfiguration on the IP addressing on your system. Asterisk is trying to send a SIP message, and passes the network packet to the operating system, and the operating system says “Whoa there buddy, that’s not right”.

However i have checked all the obvious places but my IP is correct. Also ran the update IP script.

Another find on the FreePBX forum:

So since I posted the OP, I found out what the issue was. (or at least eliminated it by changing a setting.) On old versions of FreePBX, in Advanced SIP settings, in order to enable TCP connections, you had to manually type:

tcpenable =yes
In new versions, there is now a button to enable TCP. Having both settings enabled seems to cause these issues. Most likely, the config file for sip.conf gets messed up having these similar/identical settings. Removing the old entry eliminates the errors.

This is a Vicibox 8.1.2 server with SVN 5053.
The cert was installed with vicibox-certbot.
I checked /etc/asterisk/http.conf and /etc/apache2/vhosts.d/1111-default-ssl.conf and both have the correct paths.
The webphone is working normally there are no functional problems as far as i can see.

I also see this error:

ERROR[4675]: tcptls.c:447 tcptls_stream_close: SSL_shutdown() failed: error:00000005:lib(0):func(0):DH lib, Underlying BIO error: Broken pipe

Regards, Dennis

Vicibox 9.0.1
Version: 2.14b0.5
SVN Version: 3199
DB Schema Version: 1588
Build: 200310-1801

dennis,

you ever get the broken pipe error situated?

Nox

Not actively, but i don’t think i see them anymore or i have become blind for them

I do keep seeing ‘Serious Network Trouble’ and ‘SRTCP unprotect failed because of authentication failure’

Regards, Dennis

Vicibox 9.0.1
Version: 2.14b0.5
SVN Version: 3199
DB Schema Version: 1588
Build: 200310-1801

Regards, Dennis

Vicibox 9.0.1
Version: 2.14b0.5
SVN Version: 3199
DB Schema Version: 1588
Build: 200310-1801