ValNodes

Carbon Neutral Staking - Crypto Validator backed by cutting edge innovations in active-sync, one-sign to ensure high uptime and rewards.

[ValNodes @valnodes]

@gadikian I think technical answer is no. But what happens to network if 100% of validator nodes are running pruned nodes and networks is under long/short range attack. How/who will verify Merkle tree?

[ValNodes @valnodes]

@bmc_ @zkdragon @gadikian @notionaldao @CosmWasm Oh idea was to keep state (required to verify tree) in memory and rebuild on crash from Kafka. no silver bullet here, but one with least overheads. Need to think but single node Cassandra might also work

[ValNodes @valnodes]

@zkdragon @gadikian @bmc_ @notionaldao @CosmWasm ETH started to use leveldb and so now everyone tries to fit square peg in round hole.

[ValNodes @valnodes]

@zkdragon @gadikian @bmc_ @notionaldao @CosmWasm Majority of operations are on last X blocks, and so one should use Kafka like Append only logs for those. Pruning is simple as we can drop log file beyond a time. And for other operations minor take Kafka streams and store on disks.

[ValNodes @valnodes]

@gadikian @nullMames @GoldenStaking @KevinGarrison @JunoNetwork Ha! We as community will shine, de-centralized work!

[ValNodes @valnodes]

@zmanian @Tendermint_Core @zkdragon Thanks @zmanian , appreciate it! If we can discuss over discord. Easier fix is to save ABCI Response of last height (just one block) in state.db. Rest of things for merkle tree proof to handle outside of state.db and optional for nodes that prune.

[ValNodes @valnodes]

Hola! @Tendermint_Core @ignite_com we would love your support in fixing state.db bloat by not storing ABCI Response information in state.db on every commit

Github Issue and thread on our observations -https://t.co/yCTLtgkMd8

[ValNodes @valnodes]

@mark0baricevic @Tendermint_Core @ignite_com It will get super annoying for high-frequency blocks (@EvmosOrg ), high-frequency transactions (@osmosis), heavy transaction payload (@BandProtocol), and older chains (@CosmosEcosystem). Reasons are different but state.db will bloat and cause latency and disk issues

[ValNodes @valnodes]

@lavender_five Thanks for your love! If there are things you would love to see share would do deep dive. And to fix some of this we have started to create @Tendermint_Core feature requests - https://t.co/IW51rL6QwY

[ValNodes @valnodes]

Hola! @Tendermint_Core @ignite_com we would love your support in fixing state.db bloat by not storing ABCI Response information in state.db on every commit

Github Issue and thread on our observations -https://t.co/yCTLtgkMd8

[ValNodes @valnodes]

Thanks to @zkdragon always having time to hear us out. If you like our work please support us by staking with us on Osmosis or Cosmos

https://t.co/UpgjqRmmYF

[ValNodes @valnodes]

At the end of bonding period, high volume or mature chains like already has bloated state.db and initial ABCI Information on higher level so we can't reclaim space, in turning increasing latency and having PBs of disk across all chains and validators https://t.co/yFI8I77bk8

[ValNodes @valnodes]

1/ Sloths are faster than pruning on @cosmossdk (on @LevelDB ) based networks. While pruning was introduced to reclaim space let’s see how validators will not be able to do that for months after enabling pruning. Let's see how - @tendermintHQ @Tendermint_Core

[ValNodes @valnodes]

Love this! Especially specific DB for specific use-case… That is the only way forward

[Michael Kong @michaelfkong]

8/ LevelDb => PebbleDB on RPC nodes. So far reducing I/Os by ~90% during testing.

[ValNodes @valnodes]

@mark0baricevic That requires lot of thinking as they still carry slashing risk. We are small validators who code and contribute.. and at bottom of list but want to make system fair. What is fair here, now and 5 years later when inflation is lower

[ValNodes @valnodes]

@RobinhoodApp has started to offer COMP, MATIC, SHIB, and SOL tokens from today. Safe to assume they will offer @osmosis and via them all @cosmos token by fall. At that point, will they become largest validator via custodial staking.

[ValNodes @valnodes]

Day 3: We have whale delegation alerts on @cosmos, over last 30 days avg delegations was 248K ATOMs but in last 24 hours there were delegation worth 1.15M ATOMS and one single whale delegation of 600K ATOMs (14M USD). Congratulations @HuobiGlobal

https://t.co/dR903VM9Uo

[ValNodes @valnodes]

Day 2, after expansion of @cosmoshub validator list, we had micro-delegations today 7,752 delegation events worth 106,335 ATOMs delegated. And Top 5 Validators got 52,106 delegations. Will you like to see undelegations stats or anything else? https://t.co/GlGiplZQV9

[ValNodes @valnodes]

Day 2, after expansion of @cosmoshub validator list, we had micro-delegations today 7,752 delegation events worth 106,335 ATOMs delegated. And Top 5 Validators got 52,106 delegations. Will you like to see undelegations stats or anything else? https://t.co/GlGiplZQV9

[ValNodes @valnodes]

It is amazing, how @cosmoshub remains decentralized and by expanding validator set responsibly it is even more decentralized.
In last 12 hrs, 3332 delegations worth 123,856 ATOMS but 57,221 went to @binance and ~100,000 ATOMS to top 15 validators.

[ValNodes @valnodes]

@gadikian @cosmos @CockroachDB @joe__bowman @JoeAbbey @mark0baricevic @gadikian Review and merge the pull. Basic implementation for stats in Pebble Db - https://t.co/s6CBxDQi3v

[ValNodes @valnodes]

@gadikian Yes - We should rule out all DBs using CGO and also Go version of RocksDb as it is many versions behind actual RocksDB. Does not support compaction priority and so on.

[ValNodes @valnodes]

@gadikian hmm... thinking out loud.. Should that even be allowed if one is active validator.. Obviously very different discussion.