[OpenSIPS-Users] Valid tls connection closed on a "dos" simulation
saioa.perurena at enigmedia.es
Wed Jun 1 15:06:24 CEST 2016
I've tried Opensips 2.1.3 and it seems that the problem is solved.
Thank you very much.
On 01/06/16 13:18, Liviu Chircu wrote:
> Hi Saioa,
> We have addressed this issue somewhere between OpenSIPS 1.11.6 - 1.11.7,
> and 2.1.2 - 2.1.3. Please update to the latest version (possibly even
> from GitHub , ), and let us know if it solved your problem!
> : https://github.com/OpenSIPS/opensips/tree/1.11
> : https://github.com/OpenSIPS/opensips/tree/2.1
> Liviu Chircu
> OpenSIPS Developer
> On 01.06.2016 13:22, Saioa Perurena wrote:
>> We have an opensips 1.11.6 server with tls and we simulate a dos attack
>> sending an invalid request to the tls port every 10 seconds (echo
>> "giberish" | nc sip-service.example.com 5061 ).
>> We have only two UAC connected correctly through tls, when one of this
>> clients sends an INVITE request when the dos attack is working, then
>> servers close the tls connection of that client with error:
>> ERROR:proto_tls:tls_print_errstack: TLS errstack: error:1408F10B:SSL
>> routines:SSL3_GET_RECORD:wrong version number
>> When client sends MESSAGE or OPTIONS request it does not happen.
>> If we stop the dos attack all works correctly. We can reproduce it so
>> easily, also with Opensips 2.1 version.
>> Any idea of what is happening?? Maybe it is a bug on tls? Any suggestion
>> or idea is welcome.
>> Thanks in advance.
>> Users mailing list
>> Users at lists.opensips.org
> Users mailing list
> Users at lists.opensips.org
More information about the Users