And help with what, exactly? Kamailio runs properly on a docker.

On Thu, 12 Sep 2019 at 19:05, Alex Balashov abalashov@evaristesys.com wrote:

Why?

— Sent from mobile, with due apologies for brevity and errors.

On Sep 12, 2019, at 2:00 PM, Gaurav Bmotra saigauravmehra91@gmail.com wrote:

hi i want to run kamailio in kubernetes deployment and i want to run this kamailio/kamailio-ci:5.2 image in to the kubernetse .. can anyone help

thank you

--

*Regards:* Gaurav Kumar

---

Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users

---

Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users

On Fri, Sep 13, 2019 at 08:39:23AM +0530, Gaurav Bmotra wrote:

Hi Thank you for your response Yes I m looking for HA AND scalability

If I may suggest:

Kubernetes is a very complicated solution that only works well for very particular kinds of services and scenarios, and is often not the appropriate answer for scaling out Kamailio or making it redundant. It's designed primarily for the web tier of a technology stack.

Moreover, Kamailio most often plays the role of a high-throughput centralised gateway and doesn't require much scaling, but instead is often itself used to scale out other, more limited and resource-intensive services. In a sense, Kamailio is the Kubernetes of VoIP, though obviously the analogy is limited.

Unless you've discovered a very specific reason to use Kubernetes for your particular scenario, you almost certainly do not need or want it. If "HA and scalability" are your goals, Kubernetes is very unlikely to be the answer, and if Kubernetes is the answer, I haven't seen from you a compelling explanation as to why. If you don't have one, Kubernetes is not the answer.

Making Kamailio redundant and highly available is generally accomplished through some combination of:

- Database sharing (e.g. usrloc);

- Internal replication and state sharing (e.g. DMQ);

- Heartbeat/Pacemaker/Corosync/VRRP/UCARP and appropriate network design;

- Shared-nothing or shared-little architecture;

-- Alex