Revisit the usability of DB validation

[ArturWieczorek]

This is an enhancement ticket.

Description:

I had a node (setup: just a local node running without using local byron proxy) with more than 100 epochs. It was stopped and started on the very next day.
After starting node again I received message about my database being corrupted:

[artur-Vi:cardano.node.ChainDB:Info:4] [2020-01-21 09:30:08.43 UTC] messages elided so far = 14300
ImmDbFailure (ChecksumMismatchError (Block (SlotNo {unSlotNo = 1288739})) (CRC 4256058340) (CRC 3358730171) 00059.epoch [("checkChecksum",SrcLoc {srcLocPackage = "ouroboros-consensus-0.1.0.0-Fa0PB1fvSg5HzftQ7nVswy", srcLocModule = "Ouroboros.Storage.ImmutableDB.Impl.Iterator", srcLocFile = "src/Ouroboros/Storage/ImmutableDB/Impl/Iterator.hs", srcLocStartLine = 400, srcLocStartCol = 9, srcLocEndLine = 400, srcLocEndCol = 69})])
cardano-node: The database got corrupted, please restart with validation mode enabled

As it was explained, the validation process for all epochs needs to be initiated manually by end user:

By default, the ChainDB will only check the last epoch for corruption. If you pass the --validate-db flag, then it will check all epoch files.

from end user perspective this behaviour could be improved in the future:

Duncan Coutts
If this happens on an end user machine, we cannot rely on user intervention, reading the log files and running a command. We'll have to figure out a way to do this automatically.

[mrBliss]

IntersectMBO/ouroboros-network#1551 is about this.

[mrBliss]

Now IntersectMBO/ouroboros-network#1623 is merged, the DB will automatically validate if necessary when it restarts. No user interaction is required.