[R4R]use more aggressive write cache policy

[unclezoro]

Description

Fix validator failed to produce a block occasionally.

Rationale

The issue is caused by the persistence action, it takes around 48 seconds, so the validator do not have time to produce the block.

 t=2021-05-17T07:49:55+0000 lvl=info msg="Imported new chain segment"            blocks=1   txs=517   mgas=46.932   elapsed=218.339ms   mgasps=214.950  number=7,482,287 hash=0x2a7769438c9d64ce5192        41e99414e75890a37e568db6b24b7daf5dc3814158e8 dirty="2.00 GiB"

t=2021-05-17T07:50:41+0000 lvl=info msg="Persisted trie from memory database"   nodes=5,205,776 size="1.10 GiB"   time=43.360182272s gcnodes=47,892,102 gcsize="16.62 GiB" gctime=2m53.758774587s l        ivenodes=1,017,189 livesize="323.76 MiB"

t=2021-05-17T07:50:41+0000 lvl=info msg="Imported new chain segment"            blocks=1   txs=360   mgas=34.111   elapsed=43.523s     mgasps=0.784    number=7,482,288 hash=0x8aee890ea8e6ef39515c        e2f9cfc91966e04250c019a8f180f54ac4fe27613cfc dirty="439.11 MiB"

We change the distance logic, so that there is more time for a validator to persist cache.

Example

Changes

Notable changes:

• add each change in a bullet point here
• ...

Preflight checks

• build passed (make build)
• tests passed (make test)
• manual transaction test passed

Already reviewed by

...

Related issues

... reference related issue #'s here ...

[wellttllew]

@guagualvcha

The go-ethereum team has already fixed this race condition issue in ethereum/go-ethereum@28d30b5