Please always give some additional information, like used software versions, other components involved
(like what is the remote site using which is trying to dial you), which protocolls are used, how does the
dial plan and call flow looks like, ......
I would say there are more dimensions to this.
The proper deployment should look like as Guy said, multiple SRV records pointing to specific A
records which have only one IP assigned to them. So you should fix that first anyhow and sure
if it fixes it its great.
The behavior how the lookup and handling is done is also part of the remote infrastructure trying to
dial you.
In the case of the VCS would at least not point your issue towards the single A record with two IPs (RR / RoundRobin).
I would still expect the VCS to place the call to only one VCS via SIP.
I tested that in my lab (VCS-E X7.2.2 and I could confirm, first call attempt hits only the first vcs,
a second second call attempt then hits the second VCS, but not at the same time.)
What I could see is that I got one attempt coming in via sip and h323 on the other one due to interworking
on my initiating VCS. So thats at least not wrong, but not SIP and only happened as the alias was not found.
Sure, a loop is not good and shall be prevented, but in general with the VCS
loop protection it should not break the call.
My best guess is that you have some other issues in your dialplan, zones, searchrules, transforms, ...
This will be to complex for a forum discussion as a small detail might get lost which brings in issues
in your deployment or does not really fit your requirements.
What do you think of asking your Cisco partner or some consultant to analyze your needs and reflect that to your deployment.
Please remember to rate helpful responses and identify helpful or correct answers.
