[OpenSIPS-Users] Load Balancer cluster Opensips 3.0

Mon Jun 3 07:13:57 EDT 2019

Please keep the list CC'ed to the discussion.

Thanks for the info, it seems to be an OpenSIPS bug - the anycast token 
is missing in the locally generated requests (like this OPTIONS). Could 
you open a bug report on the github tracker please ?

Best regards,

Bogdan-Andrei Iancu

OpenSIPS Founder and Developer
   https://www.opensips-solutions.com
OpenSIPS Summit 2019
   https://www.opensips.org/events/Summit-2019Amsterdam/

On 06/03/2019 02:02 PM, Антон Ершов wrote:
> Hi Bogdan,
> node1 send
>
> 2019/06/03 16:01:17.484626 10.50.0.1:5060 <http://10.50.0.1:5060> -> 
> 10.7.0.201:5070 <http://10.7.0.201:5070>
> OPTIONS sip:10.7.0.201:5070 <http://10.7.0.201:5070> SIP/2.0
> Via: SIP/2.0/UDP 10.50.0.1:5060;branch=z9hG4bK119b.69e9c536.0
> To: sip:10.7.0.201:5070 <http://10.7.0.201:5070>
> From: <sip:hchecker at 10.50.0.1 
> <mailto:sip%3Ahchecker at 10.50.0.1>>;tag=7f64a078bf34d12f31433ecf357bd91b-968b
> CSeq: 10 OPTIONS
> Call-ID: 0fa7a2ed14e0d2fe-14840 at 10.50.0.1 
> <mailto:0fa7a2ed14e0d2fe-14840 at 10.50.0.1>
> Max-Forwards: 70
> Content-Length: 0
> User-Agent: OpenSIPS (3.0.0 (x86_64/linux))
>
> and does not receive an answer as it goes to the second node
>
> node 2 send and receive answer
>
> 2019/06/03 16:01:26.320354 10.50.0.1:5060 <http://10.50.0.1:5060> -> 
> 10.7.0.201:5070 <http://10.7.0.201:5070>
> OPTIONS sip:10.7.0.201:5070 <http://10.7.0.201:5070> SIP/2.0
> Via: SIP/2.0/UDP 10.50.0.1:5060;branch=z9hG4bK894b.277e074.0
> To: sip:10.7.0.201:5070 <http://10.7.0.201:5070>
> From: <sip:hchecker at 10.50.0.1 
> <mailto:sip%3Ahchecker at 10.50.0.1>>;tag=7f64a078bf34d12f31433ecf357bd91b-9e0e
> CSeq: 10 OPTIONS
> Call-ID: 73aebff26b2035c7-17475 at 10.50.0.1 
> <mailto:73aebff26b2035c7-17475 at 10.50.0.1>
> Max-Forwards: 70
> Content-Length: 0
> User-Agent: OpenSIPS (3.0.0 (x86_64/linux))
>
>
> 2019/06/03 15:52:26.490325 10.7.0.201:5070 <http://10.7.0.201:5070> -> 
> 10.50.0.1:5060 <http://10.50.0.1:5060>
> SIP/2.0 200 OK
> Via: SIP/2.0/UDP 
> 10.50.0.1:5060;rport=5060;received=10.50.0.1;branch=z9hG4bK894b.277e074.0
> Call-ID: 73aebff26b2035c7-17475 at 10.50.0.1 
> <mailto:73aebff26b2035c7-17475 at 10.50.0.1>
> From: <sip:hchecker at 10.50.0.1 
> <mailto:sip%3Ahchecker at 10.50.0.1>>;tag=7f64a078bf34d12f31433ecf357bd91b-9e0e
> To: <sip:10.7.0.201>;tag=z9hG4bK894b.277e074.0
> CSeq: 10 OPTIONS
> Accept: application/dialog-info+xml, application/xpidf+xml, 
> application/cpim-pidf+xml, application/simple-message-summary, 
> application/pidf+xml, application/pidf+xml, 
> application/dialog-info+xml, application/simple-message-summary, 
> application/sdp, message/sipfrag;version=2.0
> Allow: OPTIONS, REGISTER, SUBSCRIBE, NOTIFY, PUBLISH, INVITE, ACK, 
> BYE, CANCEL, UPDATE, PRACK, MESSAGE, REFER
> Supported: 100rel, timer, replaces, norefersub
> Accept-Encoding: text/plain
> Accept-Language: en
> Server: Asterisk PBX 16.1.0
> Content-Length:  0
>
> my config node 1
>
> loadmodule "load_balancer.so"
> modparam("load_balancer", "db_url", 
>  "postgres://postgres:1q2w3e4r@10.3.0.1/opensips2 
> <http://postgres:1q2w3e4r@10.3.0.1/opensips2>")
> modparam("load_balancer", "probing_reply_codes", "501, 503, 403, 404")
> modparam("load_balancer", "cluster_id", 1)
> modparam("load_balancer", "cluster_sharing_tag", "node_SKN")
> modparam("load_balancer", "probing_from", "sip:hchecker at 10.50.0.1 
> <mailto:sip%3Ahchecker at 10.50.0.1>")
> modparam("load_balancer", "probing_verbose", 1)
>
> and node 2
>
> loadmodule "load_balancer.so"
> modparam("load_balancer", "db_url", 
>  "postgres://postgres:1q2w3e4r@10.3.0.1/opensips2 
> <http://postgres:1q2w3e4r@10.3.0.1/opensips2>")
> modparam("load_balancer", "probing_reply_codes", "501, 503, 403, 404")
> modparam("load_balancer", "cluster_id", 1)
> modparam("load_balancer", "cluster_sharing_tag", "node_PRM")
> modparam("load_balancer", "probing_from", "sip:hchecker at 10.50.0.1 
> <mailto:sip%3Ahchecker at 10.50.0.1>")
> modparam("load_balancer", "probing_verbose", 1)
>
>
> maybe I misunderstood how it should work?
>
>
>
> пн, 3 июн. 2019 г. в 10:53, Bogdan-Andrei Iancu <bogdan at opensips.org 
> <mailto:bogdan at opensips.org>>:
>
>     Hi Anton,
>
>     That is an interesting combination - pinging via an anycast
>     interface.
>
>     Could you paste here the OPTIONS and its reply, to check if it is
>     correctly built from the anycast perspective ?
>
>     Regards,
>
>     Bogdan-Andrei Iancu
>
>     OpenSIPS Founder and Developer
>        https://www.opensips-solutions.com
>     OpenSIPS Summit 2019
>        https://www.opensips.org/events/Summit-2019Amsterdam/
>
>     On 05/27/2019 03:16 PM, Антон Ершов wrote:
>>
>>     Hello friends!
>>
>>     I'm testing new features of opensips 3. And I discovered the
>>     following situation.
>>
>>     Configured cluster module in Anycast Сonfiguration. Load balanser
>>     module
>>
>>     has cluster_sharing_tag. The lb module sends "options" to the
>>     host from the common address, but the answer goes to the nearest
>>     opensips node of the cluster. From what the remote host in lb is
>>     considered inaccessible.
>>
>>     Something is not configured or I did not understand correctly how
>>     the load balanser should work in a cluster
>>
>>
>>
>>     _______________________________________________
>>     Users mailing list
>>     Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>>     http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20190603/9a01d976/attachment-0001.html>