We present observations of the interacting transient SN 2009ip, from the start of the outburst in October 2012 until the end of the 2012 observing season. The transient reached a peak of $M_V$=-17.7 mag before fading rapidly, with a total integrated luminosity of 1.9$\times10^{49}$ erg over the period of August-December 2012. The optical and near infrared spectra are dominated by narrow emission lines, signaling a dense circumstellar environment, together with multiple components of broad emission and absorption in H and He at velocities between 0.5-1.2$\times10^4$ km s$^{-1}$\. We see no evidence for nucleosynthesized material in SN 2009ip, even in late-time pseudo-nebular spectra. We set a limit of $<$0.02 M$_{\odot}$\ on the mass of any synthesized $^{56}$Ni from the late time lightcurve. A simple model for the narrow Balmer lines is presented, and used to derive number densities for the circumstellar medium of between $\sim 10^{9}-10^{10}$ cm$^{-3}$. Our near-infrared data does not show any excess at longer wavelengths. Our last data, taken in December 2012, shows that SN 2009ip has spectroscopically evolved to something quite similar to its appearance in late 2009, albeit with higher velocities. It is possible that neither of the eruptive and high luminosity events of SN 2009ip were induced by a core-collapse. We show that the peak and total integrated luminosity can be due to the efficient conversion of kinetic energy from colliding ejecta, and that around 0.05-0.1 M$_{\odot}$\ of material moving at 0.5-1$\times10^4$ km s$^{-1}$\ could comfortably produce the observed luminosity. The ejection of multiple shells, lack of evidence for nucleosynthesied elements and broad nebular lines, are all consistent with the pulsational-pair instability scenario. In this case the progenitor star may still exist, and will be observed after the current outburst fades.