OKNXT
Member
Offline
Activity: 70
Merit: 10
|
|
December 24, 2013, 04:40:32 AM |
|
As we know, hallmark has released. So the overall network are filled with nodes with hallmark and nodes without hallmark. Once choice is to let all the nodes remove hallmark, and the other choice is to let all those nodes add hallmark. But now we are decentralized system, nobody has that right to do that, and I don't think all the system make the same step. So a better system must be proposed and implemented to get things done.
BTW, the vote system should be released ASAP. We need to vote to make consensus.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 04:43:14 AM |
|
these forks are getting so frustrating. I know the ddos is messing us up but its next to impossible to get anywhere on the blockchain tonight.
I killed the 5 VPS I had set up seeing as nothing is working. I will gladly set them up again if we determine this to be useful.
I tried deleting the blocks and transactions and then I'm stuck on block 1. we are frozen.
hope 0.4.XX is the answer
stuck at 23421
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 04:44:03 AM |
|
Yes, I followed your thought. On the other hand, apart from removing hallmark, what else we can do to improve it so as to resistant to DDoS? A dynamic self adjustment hallmark score based on the network?
The hallmark code needs to be removed. If everyone starts removing their hallmark, the network will become more and more dependent on fewer and fewer nodes. This makes the network more susceptible to attack. The hallmarking code needs to be removed and a mandatory upgrade needs to take place. The new code needs to stipulate that hallmarked nodes that don't upgrade will be ignored or given equal weight. My other thought was increasing the "maxNumberOfConnectedPublicPeers" in 'webapps/root/WEB-INF/web.xml'. Currently, the default value is 20. Increasing this number should allow you to connect to more peers which strengthens the network.\ "A dynamic self-adjusting hallmark score based on the network" I'm not sure what you mean. Can you elaborate?
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 04:46:22 AM |
|
these forks are getting so frustrating. I know the ddos is messing us up but its next to impossible to get anywhere on the blockchain tonight.
I killed the 5 VPS I had set up seeing as nothing is working. I will gladly set them up again if we determine this to be useful.
I tried deleting the blocks and transactions and then I'm stuck on block 1. we are frozen.
hope 0.4.XX is the answer
Once a node downloads a bad block, it seems to require a complete re-download of the blockchain to get back in sync.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 04:54:41 AM |
|
these forks are getting so frustrating. I know the ddos is messing us up but its next to impossible to get anywhere on the blockchain tonight.
I killed the 5 VPS I had set up seeing as nothing is working. I will gladly set them up again if we determine this to be useful.
I tried deleting the blocks and transactions and then I'm stuck on block 1. we are frozen.
hope 0.4.XX is the answer
Once a node downloads a bad block, it seems to require a complete re-download of the blockchain to get back in sync. Hey i am seeing that, but if I trash the blocks and transactions, I'm not getting ANYTHING from the network. Sits at block 1 in my nxt folder, the blocks.nxt and transactions.nxt are time stamped to the current time, but the blockchain is stuck at 3 hours ago
|
|
|
|
OKNXT
Member
Offline
Activity: 70
Merit: 10
|
|
December 24, 2013, 04:55:38 AM |
|
"A dynamic self-adjusting hallmark score based on the network" I'm not sure what you mean. Can you elaborate?
Sorry, I have no detail. But I think the hallmark weight should be dynamic adjusted according to the peer connection, peer request, and the overall network status.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:02:36 AM |
|
in my nxt folder, the blocks.nxt and transactions.nxt are time stamped to the current time, but the blockchain is stuck at 3 hours ago
when I delete the blocks and transactions, the app makes new ones when it starts, but they don't seem to be unix files, rather just generic documents
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 05:06:30 AM |
|
these forks are getting so frustrating. I know the ddos is messing us up but its next to impossible to get anywhere on the blockchain tonight.
I killed the 5 VPS I had set up seeing as nothing is working. I will gladly set them up again if we determine this to be useful.
I tried deleting the blocks and transactions and then I'm stuck on block 1. we are frozen.
hope 0.4.XX is the answer
Once a node downloads a bad block, it seems to require a complete re-download of the blockchain to get back in sync. Hey i am seeing that, but if I trash the blocks and transactions, I'm not getting ANYTHING from the network. Sits at block 1 in my nxt folder, the blocks.nxt and transactions.nxt are time stamped to the current time, but the blockchain is stuck at 3 hours ago This is probably because you are unsuccessfully attempting to download the blockchain from hallmarked nodes, but these nodes are currently under DDoS attack and cannot fulfill your request. There are other nodes, which are not hallmarked and are up-to-date, but due to the hallmark code, your node prioritize it's request to hallmarked nodes. This is why you are not connecting to other nodes which could fulfill your request.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:07:54 AM |
|
got you loud and clear.
tell me should I set up VPS and not hallmarked? I can get 5 online in 15 minutes
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 05:16:07 AM |
|
got you loud and clear.
tell me should I set up VPS and not hallmarked? I can get 5 online in 15 minutes
No. All non-hallmarked VPS should be eliminated. They create additional and unneeded work for hallmarked nodes and provide no benefit to the network. If you start up a VPS, you should hallmark it.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:17:56 AM |
|
got you loud and clear.
tell me should I set up VPS and not hallmarked? I can get 5 online in 15 minutes
No. All non-hallmarked VPS should be eliminated. They create additional and unneeded work for hallmarked nodes and provide no benefit to the network. If you start up a VPS, you should hallmark it. But how does hallmarking help? besides taking all the bots? I had a hallmarked VPS up earlier and it had 800 connections, most with 0 weights Im reseting a Hallmarked VPS with 4gb ram and 2cores. see it it does anything https://162.243.214.183:7875/ set up and catching up.
|
|
|
|
X_X_X
Member
Offline
Activity: 117
Merit: 23
|
|
December 24, 2013, 05:30:36 AM |
|
Somebody knows why http://coinmarketcap.com/ has removed NXT from their list? it was in the 4th or 5th position some hours ago.
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:32:11 AM |
|
Somebody knows why http://coinmarketcap.com/ has removed NXT from their list? it was in the 4th or 5th position some hours ago. It sounds like it was pulling its price from dgex.com API and there was some problems with that. I don't think that coinmarketcap is an editorial list, it seems like pure data.
|
|
|
|
jl777
Legendary
Offline
Activity: 1176
Merit: 1134
|
|
December 24, 2013, 05:32:38 AM |
|
Would it be possible to allocate half the network load to hallmarked servers and half to non-hallmarked servers?
If there was a way to evenly split the traffic, then wouldn't that make it so that even if all the hallmarked servers are bogged down, the smaller servers would still be there as a backup. Not sure about how the network retry code is setup, but a random bit to choose hallmark or not hallmark, after initial timeout from hallmarked server. That way if the hallmarked servers are unmolested, the network takes advantage of the (presumably) faster hallmarked servers. If they are bogged down for whatever reason, they are only then expected to carry half the load.
If there was a way to update all the network as to the current percentage of requests being answered by the hallmarked servers (doesn't have to be totally precise), then that could be used to reduce the initial request to the hallmarked servers when their percentage goes below some threshold.
Forgive me if I am way off base, just guessing as to what the code must be doing
James
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:40:21 AM |
|
well getting that hallmark running did get me back online.
any other non trivial nxt holders that would like a VPS set up, please message me. I have a good budget right now from donated coins.
send me your hallmark and I will set you up
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 05:40:34 AM |
|
got you loud and clear.
tell me should I set up VPS and not hallmarked? I can get 5 online in 15 minutes
No. All non-hallmarked VPS should be eliminated. They create additional and unneeded work for hallmarked nodes and provide no benefit to the network. If you start up a VPS, you should hallmark it. But how does hallmarking help? besides taking all the bots? I had a hallmarked VPS up earlier and it had 800 connections, most with 0 weights Im reseting a Hallmarked VPS with 4gb ram and 2cores. see it it does anything https://162.243.214.183:7875/ set up and catching up. Hallmarking helps by taking workload off of other hallmarked servers. The larger balance the hallmarked node has, the higher priority it has with other peers. This is the reason you had 800 peers with 0 weights. They were attempting to connect to a hallmarked node to download the blockchain and sync. They behave like this regardless if the hallmarked node they are attempting to download from is in sync itself.
|
|
|
|
utopianfuture
Sr. Member
Offline
Activity: 602
Merit: 268
Internet of Value
|
|
December 24, 2013, 05:42:33 AM |
|
got you loud and clear.
tell me should I set up VPS and not hallmarked? I can get 5 online in 15 minutes
No. All non-hallmarked VPS should be eliminated. They create additional and unneeded work for hallmarked nodes and provide no benefit to the network. If you start up a VPS, you should hallmark it. But how does hallmarking help? besides taking all the bots? I had a hallmarked VPS up earlier and it had 800 connections, most with 0 weights Im reseting a Hallmarked VPS with 4gb ram and 2cores. see it it does anything https://162.243.214.183:7875/ set up and catching up. If I understand 2Kool4Skewl correctly, your hallmark node can help the network by sharing a part of the attacks caused by zombie nodes. Non hall-mark node does not help since no peer sync the blockchain from a non-hallmark node and non hall-mark node can actually cause more stresses to the existing hall-mark nodes. I think 2Kool4Skewl made a very good point here that no node should bear a over-sized responsibility to keep the network running since it would become a prime target for attackers. Now we have only a few hall-mark nodes bear major responsibility of spreading the blockchain. No wonder the network crumbles when they become targets.
|
|
|
|
virtualaddict
Newbie
Offline
Activity: 56
Merit: 0
|
|
December 24, 2013, 05:43:45 AM |
|
Sending nxt: When I try to send from my account (which has plenty of nxt), I enter the recipient, the amount, and leave it as 24hrs deadline and click send and it says "confirmed" but then no nxt is deducted from my account and it seems like nothing happens. Anyone know what I'm doing wrong?
|
|
|
|
brooklynbtc
Sr. Member
Offline
Activity: 336
Merit: 250
AKA jefdiesel
|
|
December 24, 2013, 05:45:42 AM |
|
funnily enough, my local install caught up really fast, in sync with my hallmarked node up to block 23537 then just froze there again
either just lost its connection, or its another bad fork.
|
|
|
|
2Kool4Skewl (OP)
|
|
December 24, 2013, 05:46:31 AM |
|
Would it be possible to allocate half the network load to hallmarked servers and half to non-hallmarked servers?
If there was a way to evenly split the traffic, then wouldn't that make it so that even if all the hallmarked servers are bogged down, the smaller servers would still be there as a backup. Not sure about how the network retry code is setup, but a random bit to choose hallmark or not hallmark, after initial timeout from hallmarked server. That way if the hallmarked servers are unmolested, the network takes advantage of the (presumably) faster hallmarked servers. If they are bogged down for whatever reason, they are only then expected to carry half the load.
If there was a way to update all the network as to the current percentage of requests being answered by the hallmarked servers (doesn't have to be totally precise), then that could be used to reduce the initial request to the hallmarked servers when their percentage goes below some threshold.
Forgive me if I am way off base, just guessing as to what the code must be doing
James
I'm sure that would be possible, but hallmarking nodes is just a bad idea in my opinion. Instead of using hallmarked nodes to stay in sync, nodes should simply use other nodes with the lowest ping time.
|
|
|
|
|