simulations are failing on some PR's

[aleem1314]

Summary of Bug

Simulations are failing on #10714 and #9810 PRs with the following error

simulation halted due to panic on block 35
Logs to writing to /home/callum/.simapp/simulations/2021-11-25_12:50:24.log
--- FAIL: TestAppImportExport (25.18s)
panic: invariant broken: staking: delegator shares invariant
broken delegator shares invariance:
        validator.DelegatorShares: 34148342106.937500000000000000
        sum of Delegator.Shares: 34148342106.000000000000000000


        CRITICAL please submit the following transaction:
                 tx crisis invariant-broken staking delegator-shares [recovered]
        panic: invariant broken: staking: delegator shares invariant

After debugging looks like #10216 breaking delegator shares invariant.

Version

Currently on two PRs

Steps to Reproduce

---

For Admin Use

• Not duplicate issue
• Appropriate labels applied
• Appropriate contributors tagged
• Contributor assigned/self-assigned

[amaury1093]

@alexanderbez it might be related to this logic: #10254 (comment)

[alexanderbez]

Yes, totally! So the validator.TokensFromShares(delegation.Shares).TruncateInt().IsZero() case is getting triggered (good) which deletes the delegation object (good) but the shares remain.

The RemoveValidatorTokensAndShares call should be removing the shares. @aleem1314 can you by any chance run this again but also log the shares that we pass into RemoveValidatorTokensAndShares?

I'm curious if we need to remove the dust shares...

[alexanderbez]

Either we:

1. Update the invariant
2. Evaluate if we need to remove the dust shares (and why we're currently not)

[aleem1314]

Here's is simulation logs for #9810 2021-12-14_15:10:17.log.

TOKENS_FROM_SHARES validator.TokensFromShares(delegation.Shares) =  0.937500000000000000
TOKENS_FROM_SHARES_TRUNCATE validator.TokensFromShares(delegation.Shares).TruncateInt() =  0
DELEGATIONS SHARES =  0.937500000000000000
Validator =  cosmosvaloper14rmvn8ymw7mdjwyjhmkzestgkcdskvkx6ke9e5
Delegator =  cosmos15rxvlc2xaf6sjq5fjq45gfapdjsjr97jy6wmjc
BlockHeight =  32
BLOCKTIME =  2103-10-01 17:18:06 +0000 UTC

shares that we pass into RemoveValidatorTokensAndShares

RemoveValidatorTokensAndShares =  0.062500000000000000

The delegator shares invariant is breaking with the following error

CRITICAL please submit the following transaction:
                 tx crisis invariant-broken staking delegator-shares [recovered]
        panic: invariant broken: staking: delegator shares invariant
broken delegator shares invariance:
        validator.DelegatorShares: 34148342106.937500000000000000
        sum of Delegator.Shares: 34148342106.000000000000000000

And here are simulation logs for #10714 2021-12-14_15:21:14.log.

TOKENS_FROM_SHARES validator.TokensFromShares(delegation.Shares) =  0.000000000000000000
TOKENS_FROM_SHARES_TRUNCATE validator.TokensFromShares(delegation.Shares).TruncateInt() =  0
DELEGATIONS SHARES =  660271276.166666666355950772
Validator =  cosmosvaloper10gzcdq42gzkhd09uvmzhd48nke0yllaakl42jq
Delegator =  cosmos1p3yjpqn98weh9hxj590n06z08sv9hmjzmt4t75
BlockHeight =  9
BLOCKTIME =  2066-05-12 04:15:29 +0000 UTC

shares that we pass into RemoveValidatorTokensAndShares

RemoveValidatorTokensAndShares =  38839486.833333333644049228

The delegator shares invariant is breaking with the following error

CRITICAL please submit the following transaction:
                 tx crisis invariant-broken staking delegator-shares [recovered]
        panic: invariant broken: staking: delegator shares invariant
broken delegator shares invariance:
        validator.DelegatorShares: 730298878.166666666355950772
        sum of Delegator.Shares: 70027602.000000000000000000

[alexanderbez]

I can see why the first seed/sim fails, but the later is a bit confusing. May I suggest we just revert the original PR I made until we can spend time on a proper solution.

WDYT @AmauryM? We don't have the delegation removal PR in an official release yet, do we?

[amaury1093]

May I suggest we just revert the original PR I made until we can spend time on a proper solution.

SGTM 👍. Yeah it's fine, #10216 is only on master and not in any release.