Alert: chain fork caused by pre-0.8 clients dealing badly with large blocks

[Largo]

For anyone wanting to watch the race:

https://blockchain.info/de/orphaned-blocks

[1715432877]

1715432877

[1715432877]

1715432877

[1715432877]

1715432877

[deepceleron]

Block progress - bad chain height = 225453; good chain = 225445, another hour + to go before the BIG REORG!

Soon to be orphaned:

Code:

000000000000015c50b165fcdd33556f8b44800c5298943ac70b112df480c023*225430
00000000000002d2012cc1b3fc0cceb8c156f0e698db40bf4413a210eca056c3*225431
00000000000002c9e42c23f6eed2b9f9152a440610dcfaf992357c3cbff3689c*225432
00000000000001b3c862fd66689712bc1cdd44fc004c008ada6c1b51367686f2*225433
00000000000002528688a366a9f42c07089268cd8a474961c49f787cb385807e*225434
00000000000000947452b814ed008f750023e64c8fd03bb19b3865716eeb243d*225435
00000000000001a23b3f38106f210493795a1ef6b21dd559f87ada8dfc84181b*225436
000000000000000016b3a292d63f9eea6b8177b8ba8dfd621d36fd828e5cb187*225437
000000000000019881625f04eac2046512fa1e1bd7193f56f39595d802bbbd04*225438
000000000000015aab28064a4c521d6a5325ff6e251e8ca2edfdfe6cb5bf832c*225439
00000000000003264ac7a6a74ca779997ee637b30f84debcb6ba6c26e2d09f4b*225440
0000000000000372a4c37e27587255ca77a6377391c1974f13ea55ec1d3bc01b*225441
0000000000000229cd059490b7607ca4c191c52bb40405fa052c48d72bfda8b0*225442
0000000000000213874009d269b90be16ea5c1f6a83a79ae2be3813d8972a2a3*225443
00000000000003381ed79de2fa810c0e4f0e9c0cd5d307380b779e2b5ffa982e*225444
00000000000000ec0df675972dddf0c2bdb1d436064672a6f0b79ce65ffcade6*225445
00000000000000757de9173ee2f01c9f957c8aa3b4f71b901b0fa19683ab1fa1*225446
00000000000000ec24f812807f81aadb7a665ef4ab8b8aa49df5ef789cdc13ff*225447
00000000000001f9fd0b0cab82deca5f53030911b33ba7dd7ce3fd77ad2c5a55*225448
00000000000001f15726ea57dc7702407a3ebc097ed801cb3d7b0e8a450807cb*225449
000000000000026a105de544bfbe71bca0820ac3963ba132c75ec40411b5f162*225450
00000000000003bdcc4424dea94de83e9f54ca158ae3eac59c2e97c2f90a32ad*225451
0000000000000190f5944e56c52600e9d8d4d5416432e03b324bc85c51629bcd*225452
000000000000031803e492a0103154d243c9945ef78dbdba42a1ecfdb5bc5dbb*225453

New 0.7-created blocks

Code:

00000000000001c108384350f74090433e7fcf79a606b8e797f065b130575932 225430
00000000000000f2c09bd58b84a7ef33fbc038a6fed97c072643b1e61d8ab841 225431
00000000000000362b16136f026a99ed5c59c0a4dc99e60fe87c28f3220438d6 225432
0000000000000040fcdf075576b52f2f83d852e4f356936645bcb4e745c94e15 225433
0000000000000105d2e2b2d0501ad1826bbd622c044fb77c865f98fce4781106 225434
000000000000027373b3fa7b123ca461e5c5d090c9caeb24a8bbbd32e1eaaa5d 225435
000000000000016e4bbe5cac005375b3dd799839c4b430959158b645b2dbafae 225436
000000000000022500c140cbc4ba13eebeb9859b08461c7a35fb233202ec8374 225437
0000000000000307714abe653fc80e966eca3bfe94e74697b5ec611b3c23ab96 225438
00000000000002b63d8aa07404d752703d12bf8b58a7663de30c71c719f9b2ce 225439
0000000000000025fcc0c001b11e542ae73dc4d9edc9ed9d82627563a2af5d81 225440
0000000000000133ecdb94ba8017e2977aefe3e497c48c849e3f9c4c7a091564 225441
000000000000003d3e6473a13a33ed9419bf3863d08c62e2a67abae871e2fac9 225442
00000000000001a2a869f12fab167b3506407cb62f16ff8e316a1d7b7cbd9d56 225443
000000000000007b7879107d003317325df8b3772a73f7a1c179ed99cb9d04b2 225444
000000000000012966aca980b406faa35b4c7abcc896846794f2c3f6f9b63aa0 225445

Zee blockchain killer 225430 (slush / 1MB):
http://blockchain.info/block/000000000000015c50b165fcdd33556f8b44800c5298943ac70b112df480c023

Current best of new branch:
http://blockchain.info/block/000000000000012966aca980b406faa35b4c7abcc896846794f2c3f6f9b63aa0

[justusranvier]

But what I actually meant is that it is a bit surprising that the BerkeleyDB backed versions where not tested to a maximum block size (which, as I understand it, is expressed as a #DEFINE, and which, again in my understanding, was not changed between 0.7 and 0.8.)

Maybe they were tested, but the bug only appears in certain versions of BDB. Since a lot of people compile their own node software who knows how many versions of BDB operating nodes are linked against?

We probably won't know for sure until they have time to do a full postmortem.

[adamstgBit]

is it almost over fully fixed?

[thefiniteidea]

Okay SatoshiDice is on pause now. Receiving bets but not processing anything.

so now we can stop SatoshiDICE ... how convenient

See look guys, all it took was a hard fork  

Easy!

[DoomDumas]

The original post lacked info for "regular users".  Here it is:

(1) If you are a "regular user" (not a miner), the best thing is to do nothing and wait a couple hours.
(2) If you are a "regular user",  Upgrading, downgrading, whining, FUD, etc, will make no difference.  Only miners have an incentive to do anything.
(3) Regardless of who you are, your transactions are not dead, your coins are not lost.  They will just temporarily be held up.  If you sent a transaction within the last few hours, it may take a few more hours before it's sorted out.
(4) If you insist on processing transactions right now it's probably best to wait 30+ confirmations.  It's just due diligence though ... an attacker would still need a tremendous amount of mining power, quick thinking, and a victim willing to part with a lot of BTC.
(5) By tomorrow this will be in the past and everything will appear to be normal again.  If you slept through this, you'd never know that anything happend (except for the price drop).

Let me reiterate, your coins are not at risk, your transactions are not lost.  It'll just take some time for the network to "iron itself out."  Everything will be okay.

Very good info.

+1

This is it..

Nice post !

[whitenight639]

Well Fiat never was this exciting! 

[deadserious]

So just to be clear, the only lost coins were those mined using 0.8 since the fork?

[deepceleron]

So just to be clear, the only lost coins were those mined using 0.8 since the fork?

Generated coins are not mature for 120 blocks, so they were never spendable by the miner anyway. Some pools might have some sorting out to do in their accounting systems though.

[🏰 TradeFortress 🏰]

So just to be clear, the only lost coins were those mined using 0.8 since the fork?

Only if you used a pool with instant withdraw. Some possible double spends too.

[whitenight639]

Is there any way us mere mortals without minig rigs can help? if we all come online with version 0.7 will it help or just create excessive traffic?

If you're not mining there is nothing you can do to help, and there's no reason to downgrade to 0.7 if you already upgraded earlier.

I meant those of us with a half decent Graphics card can start mining on 0.7 to process the right chain, But in all probability with my set-up re-syncing the blockchain would take more traffic and by the time my client caught up I think the blockchain race would be over.

[tvbcof]

But what I actually meant is that it is a bit surprising that the BerkeleyDB backed versions where not tested to a maximum block size (which, as I understand it, is expressed as a #DEFINE, and which, again in my understanding, was not changed between 0.7 and 0.8.)

Maybe they were tested, but the bug only appears in certain versions of BDB. Since a lot of people compile their own node software who knows how many versions of BDB operating nodes are linked against?

We probably won't know for sure until they have time to do a full postmortem.

Good point.  It'll be interesting to see the postmortem.  But it does not sound like a bug per-se.  Simply a property of BerkeleyDB's implementation (and a tunable parameter I think I read somewhere.)

[deepceleron]

Is there any way us mere mortals without minig rigs can help? if we all come online with version 0.7 will it help or just create excessive traffic?

If you're not mining there is nothing you can do to help, and there's no reason to downgrade to 0.7 if you already upgraded earlier.

I meant those of us with a half decent Graphics card can start mining on 0.7 to process the right chain, But in all probability with my set-up re-syncing the blockchain would take more traffic and by the time my client caught up I think the blockchain race would be over.

Just mine for a pool like ozcoin.net getwork that is mining the new chain. us.ozco.in:8332

[zvs]

re: miners haven't incentive to do anything.  Do they really?  Only if you're mining solo... otherwise all that matters is what the pool you're mining on is using.

[DoomDumas]

LMAO - saved !  very funny

[bbit]

is it almost over fully fixed?

Highly doubt it! If anything going to take a couple of hours to fix itself.

[johnyj]

Why that 0.8 thread is still on the top?

[whitenight639]

https://bitminter.com IS now on 0.7 according to..


https://bitcointalk.org/index.php?topic=27062.msg1613942#msg1613942



I like there simple java mining client so im gonna run that, every little helps.

[dscotese]

Seems odd that 0.7 hasn't gotten a new block for 45 minutes.  Or maybe blockchain.info isn't reporting new ones?

[zvs]

re: miners haven't incentive to do anything.  Do they really?  Only if you're mining solo... otherwise all that matters is what the pool you're mining on is using.

to reply to myself

i think there would have been a lot less ppl running around if it had been made clear that all that really matters is what client the pool you're operating on is using

as a person that mines @ a pool, i'm not downgrading from 0.8

nor am i downgrading my relay node