Normally, when the immune system is exposed to a pathogen, it creates a "memory" and responds faster the next time. It continually produces a certain level of antibodies to that pathogen. Over time, in the absence of more stimulation, the response fades - faster for some pathogens than others.
It's very different when you are talking about viruses that are persistent. Some examples would include HIV, HTLV-1/2 (also retroviruses), human herpesviruses (including VZV - chicken pox / shingles), and (sometimes) hepatitis b or c (although these may not always be chronic). Because the viruses are not eliminated entirely, the immune system continues to be stimulated, and over time, the body makes better and better antibodies against them (it selects for those that work best). Titers tend to rise over time to persistent infectious agents. People with herpes simplex tend to have very severe first outbreaks, followed by outbreaks of decreasing severity and frequency over time as the body makes more and better antibodies and learns to control the infection.
Shingles occurs only (to the best of my knowledge) in unvaccinated people who were exposed to wild type varicella virus and first developed chicken pox, and later, due to aging, the immune system starts to become less effective and the virus is able to break through. I don't think we really know how the vaccine works, although I may be wrong on this. My understanding is that there are multiple possibilities. It may be that the live, attenuated virus establishes a long term infection, which continues to offer immunity. It may be that it only primes the body to react more quickly to vzv when it encounters the wild type version, which then establishes a long term infection. It may be that no long term infection occurs - although I doubt this. It's also possible both strains coexist.
I believe the VZV strain used is not replication defective, just attenuated and less pathogenic. If it were a replication defective live virus, then it would not establish any infection.
There is currently some controversy in the development of vaccines against EBV, CMV, and the herpes simplex viruses. Some argue that even live attenuated is too dangerous, although we already use it in VZV. It does seem to produce a much more robust response than VLP glycoprotein based vaccines, and the VZV vax appears quite safe.