v6.1.0 - Better mempool filters & pruning defaults

Release v6.1.0 is a fully state machine compatible patch with v6.0.0, it contains a new mempool filter for making on-chain arbitrage txs pay a higher gas price (configurable in app.toml), and changes the default pruning txs. Two other notable changes: Adds parameters for RocksDB config to work (Thanks @faddat ), and upgrade to Tendermint v0.34.15 ( Thanks @marbar3778 )

New mempool config parameter

Makes a second config parameter for a new min fee for arbitrage txs. To get this and change this parameter, regenerate your app.toml, or add the following section to your app.toml:

###############################################################################
###                      Osmosis Mempool Configuration                      ###
###############################################################################

[osmosis-mempool]
# This is the minimum gas fee any arbitrage tx should have, denominated in uosmo per gas
# Default value of ".005" then means that a tx with 1 million gas costs (.005 uosmo/gas) * 1_000_000 gas = .005 osmo
arbitrage-min-gas-fee = ".005"

It will really help with reducing misconfigured bots spamming the chain without needing fees to go to users! If this isn't added to the app.toml, it will default to your what node otherwise set their min gas price to.

Pruning settings

Fixes a bug / changes the default pruning settings. Theres been an inconsistency in the SDK generated app.toml, where the default is actually keep-recent=(3 weeks of state), keep-every 100th state, which is why disk space is so big. (The comments in app.toml say keep-recent=100 blocks, keep-every 500th state)

This changes default pruning to keep-recent=100k states (1 week worth), and keep-every=0. This should significantly lower disk usage for new nodes that fast sync or state sync. Unfortunately existing nodes won't get a disk size reduction from this, due to this issue: cosmos/cosmos-sdk#10942 . However existing default nodes will be safe across this upgrade. Also new nodes that state sync with defualt pruning should be faster at epoch, due to getting these old states off disk, which slows IAVL.

Checkout the full changelog here: https://github.com/osmosis-labs/osmosis/blob/v6.1.0/CHANGELOG.md !