What is the correct way to provide redundancy for incoming calls from a sip trunk provider? We are trying out intelepeer and they send on 5060 to our mediation server. So if that mediation server goes down then all incoming calls won't work but outgoing will still function since we have another mediation server right?

So how exactly can I increase availability by having Intelepeer send to both mediation servers?

it would be helpful if you elaborate on your design. you say you have another mediation server. Are these two mediation nodes in the same mediation pool, or are two mediation belonging to separate pools?

You could have a trunk to each mediation server from your sip trunk provider or you could place load balancers between the sip trunk provider and the mediation servers.  

Sorry about that!

I have only one pool ATM. The mediation servers are by themselves and not collocated with any other roles.

So one pool with 2 front end, 2 edge servers, and 2 mediation servers using the LHPv2.

From what I understand having the two mediation servers will provide redundancy for outbound calls, but calls that come from the SIP trunk provider are not load balanced.

So right now I have the SIP provide sending on port 5060 to one mediation server. When I tried to load balance the incoming using Citrix Netscaler it didn't work. When I say it didn't work things like BYE were not received and sometimes you could not answer an incoming call. I think it may be because the signals were not being sent to the same mediation server (since it was load balanced).

All I know is when I took Citrix Netscaler out of the picture the calls hung up properly, incoming calls worked properly, but this scenario would cause an issue if that one mediation server went down because I would no longer receive incoming calls since that is what the SIP provider was sending too.

I hope I explained that well.

I tried the load balancing but it didn't work well. Signals like BYE were being missed and I think it was because the load balancer was sending the signal to the other mediation server that didn't contain the current call?

Read the post I just posted above ^^^

Thank you both for assisting me

I'd make sure the Netscalers are the current version and the Load Balancing is configured correctly for SIP traffic (Different then http). Take a look at: http://support.citrix.com/proddocs/topic/netscaler-load-balancing-93/ns-lb-monitors-builtin-sip-con.html#ns-lb-monitors-builtin-sip-con and http://support.citrix.com/proddocs/topic/netscaler-load-balancing-92/ns-lb-commonprotocols-sip-con.html.

Also, check to see if session affinity is enabled (not sure what it is called in NetScaler but the option should be present). Session affinity basically ensures that all session traffic is always sent to the same server.

I think that might be SOURCEIP. I've checked their documentation out and have it configured that way.

But somehow my pool is messed up now. I opened another forum because I always get the:

Couldn't sign in. Error: Error Message: Redirect attempts exceeded maximum allowed limit of (3) redirects..
Error Type: RegisterException.
Deregister Reason: None.
Response Code: 301.
Response Text: Redirect request to Home Server.

and no one can sign in now. I don't know what is going on! You have any ideas?

I decided to go with Lync load balancing. I think for inbound SIP I need two seperate trunks and have a trunk go to each mediation server.

At least this is what I believe I need to do to achieve load balancing for the incoming traffic