chainparams: Bump assumed chain params #17002

MarcoFalke commented at 8:00 PM on September 30, 2019: member

As every year, reviewers get extra point when their node is running:

assumevalid=0
checkpoints=0
on non-x86_64 hardware

See https://github.com/bitcoin/bitcoin/blob/master/doc/release-process.md#before-every-major-and-minor-release for the process.

DrahtBot added the label Docs on Sep 30, 2019

DrahtBot added the label Validation on Sep 30, 2019

fanquake added this to the milestone 0.19.0 on Sep 30, 2019

in src/chainparams.cpp:158 in fa5bc48524 outdated

 154 | @@ -155,10 +155,10 @@ class CMainParams : public CChainParams {
 155 |          };
 156 |  
 157 |          chainTxData = ChainTxData{
 158 | -            // Data from rpc: getchaintxstats 4096 0000000000000000000f1c54590ee18d15ec70e68c8cd4cfbadb1b4f11697eee
 159 | -            /* nTime    */ 1550374134,
 160 | -            /* nTxCount */ 383732546,
 161 | -            /* dTxRate  */ 3.685496590998308
 162 | +            // Data from RPC: getchaintxstats 4096 00000000000000000007b083989b3d5a937c8a6d3b633130b678fb4780643811

fanquake commented at 2:42 AM on October 1, 2019:

Why are we no longer using the same block for the getchaintxstats call?

Sjors commented at 10:45 AM on October 1, 2019:

I would also prefer using the same block.

in doc/release-process.md:29 in fa5bc48524 outdated

  25 | @@ -26,7 +26,7 @@ Release Process
  26 |  * Update hardcoded [seeds](/contrib/seeds/README.md), see [this pull request](https://github.com/bitcoin/bitcoin/pull/7415) for an example.
  27 |  * Update [`src/chainparams.cpp`](/src/chainparams.cpp) m_assumed_blockchain_size and m_assumed_chain_state_size with the current size plus some overhead (see [this](#how-to-calculate-m_assumed_blockchain_size-and-m_assumed_chain_state_size) for information on how to calculate them).
  28 |  * Update `src/chainparams.cpp` chainTxData with statistics about the transaction count and rate. Use the output of the RPC `getchaintxstats`, see
  29 | -  [this pull request](https://github.com/bitcoin/bitcoin/pull/12270) for an example. Reviewers can verify the results by running `getchaintxstats <window_block_count> <window_last_block_hash>` with the `window_block_count` and `window_last_block_hash` from your output.

Sjors commented at 10:32 AM on October 1, 2019:

Did you mean to remove the instruction?

MarcoFalke force-pushed on Oct 1, 2019

chainparams: Bump assumed chain params fa3a733116

MarcoFalke force-pushed on Oct 1, 2019

MarcoFalke commented at 11:35 AM on October 1, 2019: member

Took the ChainTxData from the interval that has the assumed valid block as the last block, as requested by @Sjors and @fanquake

laanwj commented at 12:36 PM on October 1, 2019: member

ACK fa3a7331160d1a460b1c15fca1810e98070d629c

Checked on ARM32 (with assumevalid and checkpoints enabled, unfortunately)

mainnet chainTxData matches

{
  "time": 1569926786,
  "txcount": 460596047,
  …
  "txrate": 3.77848885073875
}

getblockchaininfo w/ mainnet mMinimumChainWork, work is clearly ahead

{
  "chainwork": "000000000000000000000000000000000000000008ea99c8688b9290267a49f8",
}
consensus.nMinimumChainWork = uint256S("0x000000000000000000000000000000000000000008ea3cf107ae0dec57f03fe8");

mainnet defaultAssumeValid matches expected

cli getblockhash 597379
00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf

consensus.defaultAssumeValid = uint256S("0x00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf"); // 597379

Block 597379 is 8 blocks back at this time. It should be deep enough by the time 0.19.0 is actually released,

fanquake approved

fanquake commented at 12:36 PM on October 1, 2019: member

ACK fa3a7331160d1a460b1c15fca1810e98070d629c - checked the mainnet values. I have notes on reviewing assumevalid updates in core-review.

src/bitcoin-cli getchaintxstats 4096 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf
{
  "time": 1569926786,
  "txcount": 460596047,
  "window_final_block_hash": "00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf",
  "window_final_block_height": 597379,
  "window_block_count": 4096,
  "window_tx_count": 8610103,
  "window_interval": 2278716,
  "txrate": 3.77848885073875
}

src/bitcoin-cli getblock 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf
{
  "hash": "00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf",
  "confirmations": 8,
  "strippedsize": 925201,
  "size": 1217775,
  "weight": 3993378,
  "height": 597379,
  "version": 536870912,
  "versionHex": "20000000",
  "merkleroot": "e92ac24aaef4aa1ece52cc41e975bfc83cb1f1fe93923bdb7a6bff824c50b084",
  "tx": [ <snip> ],
  "time": 1569926786,
  "mediantime": 1569922918,
  "nonce": 919954964,
  "bits": "17160f24",
  "difficulty": 12759819404408.79,
  "chainwork": "000000000000000000000000000000000000000008ea3cf107ae0dec57f03fe8",
  "nTx": 2590,
  "previousblockhash": "00000000000000000002fc5679f64f7fe636318ade80bb75f4978d0c75658eb0",
  "nextblockhash": "000000000000000000061b82466f5aef92d169f044f2ac38660ab402fe9d1de4"

<details><summary>Show signature</summary>

Signature:

</details>

fanquake referenced this in commit ca7a761117 on Oct 1, 2019

MarcoFalke commented at 1:04 PM on October 1, 2019: member

Would be nice to get an ACK on the datadir size and maybe one ACK for testnet as well.

Sjors commented at 1:41 PM on October 1, 2019: member

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

ACK fa3a7331160d1a460b1c15fca1810e98070d629c for mainnet on macOS 10.14.6.

bitcoin-cli getchaintxstats 4096 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf
{
  "time": 1569926786,
  "txcount": 460596047,
  "window_final_block_hash": "00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf",
  "window_final_block_height": 597379,
  "window_block_count": 4096,
  "window_tx_count": 8610103,
  "window_interval": 2278716,
  "txrate": 3.77848885073875
}

bitcoin-cli getblockheader 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf{
  "mediantime": 1569922918,
  "chainwork": "000000000000000000000000000000000000000008ea3cf107ae0dec57f03fe8",
}

du -sh *
257G	blocks
3,5G	chainstate
-----BEGIN PGP SIGNATURE-----
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=GsQC
-----END PGP SIGNATURE-----

I didn't run IBD from scratch on macOS. I just spun up an Orange Pi for armv7 for bonus points, but that'll take a while.

jamesob commented at 2:42 PM on October 1, 2019: member

ACK https://github.com/bitcoin/bitcoin/pull/17002/commits/fa3a7331160d1a460b1c15fca1810e98070d629c

Verifying new mainnet assumevalid value (00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf):

$ cli getblock 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf

{
  "hash": "00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf",
  "confirmations": 22,
  "strippedsize": 925201,
  "size": 1217775,
  "weight": 3993378,
  "height": 597379,
  "version": 536870912,
  "versionHex": "20000000",
  "merkleroot": "e92ac24aaef4aa1ece52cc41e975bfc83cb1f1fe93923bdb7a6bff824c50b084",
  "tx": [
     ...
  ],
  "time": 1569926786,
  "mediantime": 1569922918,
  "nonce": 919954964,
  "bits": "17160f24",
  "difficulty": 12759819404408.79,
  "chainwork": "000000000000000000000000000000000000000008ea3cf107ae0dec57f03fe8",
  "nTx": 2590,
  "previousblockhash": "00000000000000000002fc5679f64f7fe636318ade80bb75f4978d0c75658eb0",
  "nextblockhash": "000000000000000000061b82466f5aef92d169f044f2ac38660ab402fe9d1de4"
}

Verifying new mainnet chainTxData value:

$ cli getchaintxstats 4096 00000000000000000005f8920febd3925f8272a6a71237563d78c2edfdd09ddf

{
  "time": 1569926786,
  "txcount": 460596047,
  ...
  "txrate": 3.77848885073875
}

And new mainnet *_size values:

$ du -hs /data2/bitcoin/blocks

258G    /data2/bitcoin/blocks

$ du -hs /data2/bitcoin/chainstate

3.5G    /data2/bitcoin/chainstate

From a very uncool x86_64 machine synced with a very uncool default configuration.

</details>

MarcoFalke commented at 3:58 PM on October 1, 2019: member

testnet3, anyone?

Sjors commented at 5:26 PM on October 1, 2019: member

Chainwork doesn't match for me with testnet3.

$ bitcoin-cli -testnet getchaintxstats 4096 00000000000000b7ab6ce61eb6d571003fbe5fe892da4c9b740c49a07542462d
{
  "time": 1569741320,
  "txcount": 52318009,
  "window_final_block_hash": "00000000000000b7ab6ce61eb6d571003fbe5fe892da4c9b740c49a07542462d",
  "window_final_block_height": 1580000,
  "window_block_count": 4096,
  "window_tx_count": 443460,
  "window_interval": 2923265,
  "txrate": 0.1517002392872353
}
$ bitcoin-cli -testnet getblockheader 00000000000000b7ab6ce61eb6d571003fbe5fe892da4c9b740c49a07542462d
{
      "chainwork": "00000000000000000000000000000000000000000000012b233ad65161453b4e",
}

du -hs *
 30G	blocks
1,2G	chainstate

MarcoFalke commented at 5:51 PM on October 1, 2019: member

The chainwork is for the best block (not for a past block). The total work can never decrease, but a specific block might be reorged out on testnet

MarcoFalke commented at 5:56 PM on October 1, 2019: member

So you could check that the chainwork I put in is more than the chainwork of the assumed valid testnet block, but also not more than the chainwork of your current testnet chain tip.

Sjors commented at 7:41 PM on October 1, 2019: member

0x00000000000000000000000000000000000000000000012b2a... < 0x00000000000000000000000000000000000000000000012b3a... and close enough, so ACK. Though it's easier to check this if you do use the same block.

MarcoFalke commented at 7:50 PM on October 1, 2019: member

so ACK. Though it's easier to check this if you do use the same block.

Makes sense as this is only for testnet, so the exact value doesn't matter a lot anyway.

fanquake referenced this in commit 4b51ed89cf on Oct 2, 2019

fanquake merged this on Oct 2, 2019

fanquake closed this on Oct 2, 2019

MarcoFalke deleted the branch on Oct 2, 2019

Sjors commented at 2:19 PM on October 23, 2019: member

My 32 bit ARM Orange Pi Ubuntu Bionic with Armbian Linux 4.19.62-sunxi just succesfully reached block 597380; only took two weeks and a bunch of handholding. I used assumevalid=0 and checkpoints=0. I get the same results as in my macOS ACK.

DrahtBot locked this on Dec 16, 2021