So I have a cluster of FS servers using a common shared DB. There is host weighting and some loosely intelligent dialplan generation to attempt to get related dialogs started on the same instance but I still occasionally encounter the "nightmare xfer" section of sofia.c which basically guarantees this call is doomed. To be clear, this happens when for some reason an endpoint ends up on a call through FS1, then makes an outbound call that for some cacophony of reasons ends up on FS2 and then hit xfer. There isnt anything in the signaling from most endpoints that indicate to me the invite beginning the 2nd leg is implicitly related to the 1st so i cant do any clever routing at the outset. What can i do to try and make this recoverable? Since i dont even get a dialplan request for the refer im struggling with how to affect this in a way that might make the call succeed. Thanks in advance -Ryan