Removes incorrect try-state check in staking

[gpestana]

The current logic does not completely prevent the ledger active balance to remain above ED and there are code/config paths in which ledger.active may be 0. This PR removes the try-state check that is testing those invariants.

Closes #14246

[kianenigma]

Has an issue been made about minStake/minNominatorStake/minValidator stake? if we had minStake, then it would have been reasonable to at least raise a warning if ledger.active is less than minStake, although even this is not great as minStake can reduce.

[kianenigma]

Actually, I found the reason why we had this test: this code should prevent ledger.active from ever getting less than

substrate/frame/staking/src/pallet/mod.rs

Line 1011 in 1bbde5d

if ledger.active < T::Currency::minimum_balance() {

.

Although, slashing could still violate this.

I think it is still better to first check a few of the stashes that cause this, and see what path they went through.

[gpestana]

Actually, I found the reason why we had this test: this code should prevent ledger.active from ever getting less than (ED)

Although, slashing could still violate this.

I think that's correct. However, if a stash has been slashed to the point of the active/total balance being below ED, someone needs to call reap for the stash to be killed and removed. Otherwise, the entry with ledger.total and ledger.active will linger in the ledger.

I think it is still better to first check a few of the stashes that cause this, and see what path they went through.

As mentioned in #14246, I see two paths for ledger.active to be below ED:

1. First chill stash and not call withdraw_unbonded after the bonding period;
2. Slashing occurs that drops ledger.total below ED and no reap is called.

From my current understanding, the try-state consistence check that this PR is removing is not correct and should be removed.

We may want to think about a way to call withdraw_unbonded permissionlessly after a minimum of eras after the unbonding period has passed, so that we can free up space in storage (more on this in the issue above).

[kianenigma]

Fair enough.

All in all, I still find this situation with our assumptions around ledger a bit confusing. I hope you take a fundamental look at it and add new reasonable checks, if any.

[kianenigma]

bot merge

[kianenigma]

bot rebase

[paritytech-processbot]

Rebased

[gpestana]

bot merge