I have some finding about the recent transaction problem. I knew what was the problem when I fixed my test pool - transactions with size bigger than 10,000 bytes were hardly accepted in the blockchain. I took my investigation further and here are my finding:
1. Minergate uses old or modified code:
First of all, I use the latest version of dashcoind from 16 Sep on my pool. Here is the --version output:
> dashcoind --version
2014-Oct-20 10:18:48.631403 Starting...
2014-Oct-20 10:18:48.633060 dashcoin v1.0.2.358()
dashcoin v1.0.2.358()
2. Minergate started to use this code on 2014/10/06 10:28:17I looked at the blockchain for odd blocks and they started to appear here:
http://chainradar.com/dsh/blocks/2241/3067225 14-10-06 10:28:17 576 1 63d91c6d5f7f9eab1cc2ad3f3be0435fba2427694ba4579e1cdb6077e8e47110
Secondly the statement that MinerGate has moved to modified daemon code on 10/06/2014 is totally incorrect. CRYPTONOTE_BLOCK_GRANTED_FULL_REWARD_ZONE represents a
base median value of block total transactions size and can be shifted forward. As you can see the current median of 63d91c6d5f7f9eab1cc2ad3f3be0435fba2427694ba4579e1cdb6077e8e47110 '
malicious' block equals 23,142 bytes, which means it can contain up to 23 Kb of transactions. Real block size depends on the daemon's tx pool state - the mining pool have nothing to include into the block if it is empty. So, since this isn't correct, the remaining statements are incorrect also.
Lastly - for whose benefit? Please answer what my pool will gain from this potential modification?