improve robustness, hardening

[LarryRuane]

There are likely improvements needed (beyond Aditya's branch) to increase the availability of the lightwalletd service. If the zcashd node it's using restarts, for example, or is taking a long time to sync the blockchain, the lightwalletd process should not require a restart. It should patiently wait, reconnect, retry.

[LarryRuane]

I now disagree with my own comment above. If zcashd restarts, the new instance may be running a different configuration, and lightwalletd may have cached this configuration, making it now out of date. Changes should probably be made in the opposite direction: If lightwalletd has any problem communicating with zcashd, it (lightwalletd) should fail fast. There must be some kind of monitoring mechanism (if not human) that would be better positioned to decide when to restart lightwalletd.

I'll leave this issue open so we can consider making those kinds of changes in the ffuture.