missing initdb panic handler

[pmp-p]

Since we are about to integrate configuration changes from user that may happen right after initdb but without real validation, any persistent filesystem could get corrupted by wrong or missing configuration files.

We need to handle any abort() at initdb stage and be able to clear data folder so the process can start again after fixing conditions.

Actually it would eg require to go in javascript tools from browser and manually delete the indexed database.

related #415

[pmp-p]

pg_resetwal ( wasi like pg_dump ) could be inlined ( also shipped as wastime-wasi/node-wasi) to try solve the bad checkpoint case.
clearing db in last resort.