We implemented a work around for this issue by adding a couple seconds sleep between starting up the second X server.
We implemented a work around for this issue by adding a couple seconds sleep between starting up the second X server.