monocolor
|
|
September 24, 2013, 01:41:24 AM |
|
I also had a sync problem. But when I get the version 1.7 and sync'ed from the scratch (delete everything in the config dir except wallet and config files), the sync seems OK. And I sync'ed to block 247330. I am not sure if this is a forked chain or is the good chain. I looked at the block explorer in the OP, and it shows at block 246953 for about 2 days, is the block explorer stuck? or there's a fork?
Anyway, I am glad to get over the block 246953, and be in a block close to the diff retarget switch block of 248000.
|
|
|
|
Stouse49
|
|
September 24, 2013, 03:13:25 AM |
|
I looked at the block explorer in the OP, and it shows at block 246953 for about 2 days, is the block explorer stuck? or there's a fork?
In my opinion this is not the result of a fork, but a difficulty adjustment stall. At block 245000, the developer changed the difficulty retarget method. Before 245000, the code appeared to adjust the difficulty every 120 blocks, based on the block solving time of those 120 blocks. The Difficulty would adjust by a maximum of 4x (four times). After 245000, in an effort to reduce the strip mining effects by the multipools, the retarget scheme changed to adjust every block based on the block solving time of the previous 120 blocks with a maximum adjustment of 4x. This turned out to be a disaster with the most recent hash rate increase on the coin, because in the space of 10 blocks, the difficulty increased by a factor of over 1 million. Using the old method (pre 245000), the difficulty would have only increased by a factor of 4. Block Difficulty 246943 0.00024414 246944 0.0004377 246945 0.0017508 246946 0.007 246947 0.028 246948 0.112 246949 0.448 246950 1.792 246951 7.171 246952 28.685 246953 114.742 246954 249.59 Not solved The hash rate reported by the client is incorrect. The client reports 35.7 GH/s. The hashrate calculation does not account for the changing difficulty. Instead it assumes that last 120 blocks had a difficulty of 114.74, which results in a 35.7 GH/s number. The code for calculating the network hash rate is another place an alt coin needs to be changed, when cloned from Litecoin. My calculations for hash rate are as follows: Block MH/s 246953 407.6 246952 535.7 246951 1,466.6 246950 171.0 246949 62.1 246948 24.1 246947 30.1 246946 30.1 246945 4.3 246944 1.9 246943 1.0 Since these hashrates are based on individual block times, they cannot be taken to be exact network hash rates due to luck. The size of the multipool that was added may have been around 1 to 2 GH/s.
|
BTC:
|
|
|
erk
|
|
September 24, 2013, 03:24:07 AM |
|
I looked at the block explorer in the OP, and it shows at block 246953 for about 2 days, is the block explorer stuck? or there's a fork?
In my opinion this is not the result of a fork, but a difficulty adjustment stall. At block 245000, the developer changed the difficulty retarget method. Before 245000, the code appeared to adjust the difficulty every 120 blocks, based on the block solving time of those 120 blocks. The Difficulty would adjust by a maximum of 4x (four times). After 245000, in an effort to reduce the strip mining effects by the multipools, the retarget scheme changed to adjust every block based on the block solving time of the previous 120 blocks with a maximum adjustment of 4x. This turned out to be a disaster with the most recent hash rate increase on the coin, because in the space of 10 blocks, the difficulty increased by a factor of over 1 million. Using the old method (pre 245000), the difficulty would have only increased by a factor of 4. Block Difficulty 246943 0.00024414 246944 0.0004377 246945 0.0017508 246946 0.007 246947 0.028 246948 0.112 246949 0.448 246950 1.792 246951 7.171 246952 28.685 246953 114.742 246954 249.59 Not solved ... Gamecoin GME has a similar problem. You can't average your diff over a large number of block without making it overshoot and lock you at high diff. To solve it, they settled on a recalc ever 12 block with a 48 block history, even that's too much history to average over, but they only do a 10% diff change on each recalc so it compensated a bit. If you were going to do a diff recalc every block, you wouldn't want to average over the last 120blocks, more like 4 or 5 blocks would be enough. Now you have a real problem, it will probably need a new client release to fix as the next diff change will be higher not lower. Better get onto it quickly before exchanges and pools get mad. The changes will probably have to kick in on the next block, or the client will have to start at a lower diff somehow.
|
|
|
|
monocolor
|
|
September 24, 2013, 03:45:14 AM |
|
Dev already changed the diff retarget to the PPCoin algorithm, and set the switch block to be 248000. I think this algorithm should be reliable, as it has been used by all pos coins. The block 248000 may be a little conservative, but if people don't switch then we are in a bigger problem. It really depends on the diff, some cycles the diff did not go very high, but some are really high, it may depends on the hashrate at that time. It was faster a few days ago.
Some people attacked the IFC for sure. Any pow coin with 60 blocks or above retarget diff cycle can be attacked exactly the same way as IFC. That is, use high hash to boost the diff, then leave, then come back at low diff etc. After a few cycles, the coin will likely be blocked due to its high diff. IFC being the first one attacked this way possibly because its popularity.
But it looks like I am in another chain... I am at block 247330 sync'ed.
There maybe multiple chains out there, I saw one at 254488, another at 244999, but these two are not sync'ed though.
|
|
|
|
erk
|
|
September 24, 2013, 03:58:08 AM Last edit: September 24, 2013, 06:02:14 AM by erk |
|
Dev already changed the diff retarget to the PPCoin algorithm, and set the switch block to be 248000. I think this algorithm should be reliable, as it has been used by all pos coins. The block 248000 may be a little conservative, but if people don't switch then we are in a bigger problem. It really depends on the diff, some cycles the diff did not go very high, but some are really high, it may depends on the hashrate at that time. It was faster a few days ago.
Some people attacked the IFC for sure. Any pow coin with 60 blocks or above retarget diff cycle can be attacked exactly the same way as IFC. That is, use high hash to boost the diff, then leave, then come back at low diff etc. After a few cycles, the coin will likely be blocked due to its high diff. IFC being the first one attacked this way possibly because its popularity.
But it looks like I am in another chain... I am at block 247330 sync'ed.
There maybe multiple chains out there, I saw one at 254488, another at 244999, but these two are not sync'ed though.
246953 looks like the last block. http://exploretheblocks.com:2750/chain/Infinitecoin
|
|
|
|
monocolor
|
|
September 24, 2013, 05:42:12 AM |
|
Dev already changed the diff retarget to the PPCoin algorithm, and set the switch block to be 248000. I think this algorithm should be reliable, as it has been used by all pos coins. The block 248000 may be a little conservative, but if people don't switch then we are in a bigger problem. It really depends on the diff, some cycles the diff did not go very high, but some are really high, it may depends on the hashrate at that time. It was faster a few days ago.
Some people attacked the IFC for sure. Any pow coin with 60 blocks or above retarget diff cycle can be attacked exactly the same way as IFC. That is, use high hash to boost the diff, then leave, then come back at low diff etc. After a few cycles, the coin will likely be blocked due to its high diff. IFC being the first one attacked this way possibly because its popularity.
But it looks like I am in another chain... I am at block 247330 sync'ed.
There maybe multiple chains out there, I saw one at 254488, another at 244999, but these two are not sync'ed though.
246953 looks like the last block. If you sync from scratch with version 1.7, I think you will sync to this chain, which is currently at block 247332, this looks the correct chain to me. I tried it in another computer, I got the same thing.
|
|
|
|
fisheater (OP)
|
|
September 24, 2013, 07:53:10 AM |
|
Dev already changed the diff retarget to the PPCoin algorithm, and set the switch block to be 248000. I think this algorithm should be reliable, as it has been used by all pos coins. The block 248000 may be a little conservative, but if people don't switch then we are in a bigger problem. It really depends on the diff, some cycles the diff did not go very high, but some are really high, it may depends on the hashrate at that time. It was faster a few days ago.
Some people attacked the IFC for sure. Any pow coin with 60 blocks or above retarget diff cycle can be attacked exactly the same way as IFC. That is, use high hash to boost the diff, then leave, then come back at low diff etc. After a few cycles, the coin will likely be blocked due to its high diff. IFC being the first one attacked this way possibly because its popularity.
But it looks like I am in another chain... I am at block 247330 sync'ed.
There maybe multiple chains out there, I saw one at 254488, another at 244999, but these two are not sync'ed though.
246953 looks like the last block. If you sync from scratch with version 1.7, I think you will sync to this chain, which is currently at block 247332, this looks the correct chain to me. I tried it in another computer, I got the same thing. Looks like this is a forked blockchain that gets around that block that has huge diff, which virtually jammed the chain. Let me see if I can sync to it.
|
|
|
|
erk
|
|
September 24, 2013, 08:41:16 AM |
|
Dev already changed the diff retarget to the PPCoin algorithm, and set the switch block to be 248000. I think this algorithm should be reliable, as it has been used by all pos coins. The block 248000 may be a little conservative, but if people don't switch then we are in a bigger problem. It really depends on the diff, some cycles the diff did not go very high, but some are really high, it may depends on the hashrate at that time. It was faster a few days ago.
Some people attacked the IFC for sure. Any pow coin with 60 blocks or above retarget diff cycle can be attacked exactly the same way as IFC. That is, use high hash to boost the diff, then leave, then come back at low diff etc. After a few cycles, the coin will likely be blocked due to its high diff. IFC being the first one attacked this way possibly because its popularity.
But it looks like I am in another chain... I am at block 247330 sync'ed.
There maybe multiple chains out there, I saw one at 254488, another at 244999, but these two are not sync'ed though.
246953 looks like the last block. If you sync from scratch with version 1.7, I think you will sync to this chain, which is currently at block 247332, this looks the correct chain to me. I tried it in another computer, I got the same thing. Looks like this is a forked blockchain that gets around that block that has huge diff, which virtually jammed the chain. Let me see if I can sync to it. Trouble is that the shorter chain is jammed on Cryptsy at high diff with transaction confirmations riding on it.
|
|
|
|
monocolor
|
|
September 24, 2013, 08:57:53 AM |
|
Looks like the chain I sync'ed is a forked chain at block 246948, so it does not violate any checkpoints. If some confirmations are overwritten, will anyone lose coins? Or the transactions will be considered not happened?
If someone will lose big coins, then we better wait. Otherwise we may use the forked chain which will get to the switch point earlier. The problem with block 246953 is that it is at big diff, no idea when it will be solved.
Apparently there's at least one more blockchain at block 254501 now.
|
|
|
|
markn
Member
Offline
Activity: 81
Merit: 10
|
|
September 24, 2013, 09:02:33 AM |
|
Looks like the chain I sync'ed is a forked chain at block 146948, so it does not violate any checkpoints. If some confirmations are overwritten, will anyone lose coins? Or the transactions will be considered not happened?
If someone will lose big coins, then we better wait. Otherwise we may use the forked chain which will get to the switch point earlier. The problem with block 146953 is that it is at big diff, no idea when it will be solved.
Thank you for issue rectification and quick help!
|
|
|
|
c1010010
|
|
September 24, 2013, 03:32:44 PM |
|
So how long are we going to have to trudge through these mega-difficult blocks?
|
|
|
|
iPaulito
Full Member
Offline
Activity: 150
Merit: 100
1EDwkxCjCMGGNQqZdxa8FwheMHXSoQe4TU
|
|
September 24, 2013, 07:05:14 PM |
|
Somebody just dumped 10 Bilions IFC which is 11% of existence. Any ideas why?
|
|
|
|
erk
|
|
September 24, 2013, 08:45:55 PM |
|
Somebody just dumped 10 Bilions IFC which is 11% of existence. Any ideas why?
It wasn't on the main block chain, that hasn't advanced in days therefore no transactions.
|
|
|
|
zackclark70
Legendary
Offline
Activity: 868
Merit: 1000
ADT developer
|
|
September 24, 2013, 08:49:32 PM |
|
Somebody just dumped 10 Bilions IFC which is 11% of existence. Any ideas why?
It wasn't on the main block chain, that hasn't advanced in days therefore no transactions. dumped on cryptsy so i am guessing that it was already on the exchange
|
|
|
|
erk
|
|
September 24, 2013, 08:55:14 PM |
|
Somebody just dumped 10 Bilions IFC which is 11% of existence. Any ideas why?
It wasn't on the main block chain, that hasn't advanced in days therefore no transactions. dumped on cryptsy so i am guessing that it was already on the exchange I have an IFC deposit on Cryptsy stuck on 5 confirms for days now, I will not be the only one, so I think they might freeze the coin, who knows they might reverse the dump, but whoever bought the coins wont be able to withdraw them so it doesn't really matter. The will just be stuck on Cryptsy forever if the block chain doesn't get fixed.
|
|
|
|
c1010010
|
|
September 24, 2013, 09:44:01 PM |
|
Well, that seems like a problem, no?
|
|
|
|
CaptChadd
|
|
September 24, 2013, 10:03:51 PM |
|
I bet right now they are forks in this chain all over the place. Everyone needs to choose an update and stick with that one, even if it means going back to basics with some of the code.
|
|
|
|
monocolor
|
|
September 25, 2013, 03:02:18 AM Last edit: September 25, 2013, 07:47:43 AM by monocolor |
|
OK I think this is enough. For 3 days, I've been waiting for this super hard block to be solved, and it was not solved, and I don't know how long it will take. The blockchain is jammed. The coin is about to be dead. Thanks to all the attackers to put IFC in this situation. Luckily we have a way to get out of it. I was sync'ed to an apparent forked chain, and I watched it go from block 247330 to block 248722. I started to mine it sometimes ago, it works fine. With version 1.7 and starting from 248000, the diff retarget becomes normal and everything seems working. I will not be stuck at block 246953 in the old chain forever. I am using the new chain. By comparing the forked chain with the one in block explorer, I see it forked out of block 246948, 5 blocks before the one we get stuck. So it is not too bad, only 5 blocks will be lost. You can use 1.7 client and sync from scratch (you need to clean up the config dir, NOT delete wallet and config file, then start to sync with your client), you should get to the forked chain like I did. The sync will take about 3 hours to complete. Or you can download the chain here (pretty big: 228M) : https://mega.co.nz/#!ZghiRIzT!GlZMiA2_Nv7MQ6V4QqwJRzgfSI5IgkKsnvnZbzJ2Pn4 unrar the file, and copy the files into your config dir (please backup your existing chain, so if you don't like it you can go back). Also here are my peers, that you may want to use as addnode. 66.87.67.255 108.62.211.7 66.63.176.231 80.7.165.182 24.144.220.244 192.241.222.102 98.18.252.73 192.249.59.81 75.152.93.200 95.170.82.34 67.177.3.56 80.234.71.58 Note, apparently there are more than one forked chains, but other chains can not be sync up, possibly they violate some constraints such as checkpoints. Once you sync to 2487xx, you should be able to start mining with 1.7 client.
|
|
|
|
tokyoghetto
Legendary
Offline
Activity: 1232
Merit: 1000
|
|
September 25, 2013, 03:10:16 AM |
|
I like monocolors solution.
|
|
|
|
wmikrut
|
|
September 25, 2013, 05:04:24 AM |
|
Excellent solution monocolor!
I have been stuck for a while now... I loaded your block chain, put in the listed nodes on the new client and was 100% synced in minutes. Thank you!
|
I will NEVER ask for any kind of funds up front in a buy/sale of anything on bitcointalk.
BM-2cTFihJKmSwusMAoYuUHPvpx56Jozv64KK
|
|
|
|