• Forum has been upgraded, all links, images, etc are as they were. Please see Official Announcements for more information

v0.10.15.x Testing

Status
Not open for further replies.
Unable to get a connection to testnet, running the test wallet, do I need to chuck an IP address in somewhere? Windows wallet...
 
Should I stop these nodes from collecting on mainnet? It can't be good for the network. Looks like I'm up to four(*nope, three) payments. Definitely do not have multiple users running darkcoind on the same instance. I haven't touched anything in case we need to gather data on them. Looks like at least one is not showing up in testnet as active tMN anymore because of version number. I can update the version on tnet so that the IP shows up in both testnet and mainnet. Just haven't done so because this is so weird and I don't want to interrupt it if we need data.

*Anyway, I know Evan has a lot on his plate. I'll just not touch them till I hear different.
I'm confused, I thought you said later that you were mistaken and that your friend tried as well and errored out. I find it hard to believe you have a testnet private key working on mainnet with functionally, they begin with different characters and are using two different ports.
 
I'm confused, I thought you said later that you were mistaken and that your friend tried as well and errored out. I find it hard to believe you have a testnet private key working on mainnet with functionally, they begin with different characters and are using two different ports.
I woudln't believe me either, which I'm sure *is why this is not getting much attention. At first I thought I had left the keys the same, but on closer inspection it looks like I did in fact change the keys after changing to testnet. I have both port 9999 open and 19999. When I ask darkcoind getinfo both state that they are on testnet, one is v .15.1 the other is a version .14. THey are both showing as active on mainnet and receiving payments. Someone else tried to replicate what I did, left the keys the same, and it rejected them. That is as far as they got in trying to replicate this though, as far as I know.

* I believe the only reason they are showing active on mainnet is the network memory, not the keys.
 
Last edited by a moderator:
It's looping through this:

2014-10-11 04:27:53 Misbehaving: 54.189.202.6:9999 (0 -> 1)
2014-10-11 04:27:53 ProcessMessage(dseep, 116 bytes) FAILED
2014-10-11 04:27:53 partner 54.189.202.6:9999 using obsolete version 70038; disconnecting
2014-10-11 04:27:53 ProcessMessage(version, 104 bytes) FAILED
2014-10-11 04:27:53 disconnecting node 54.189.202.6:9999
2014-10-11 04:27:54 trying connection 192.3.172.233:9999 lastseen=5.7days
2014-10-11 04:27:54 connected 192.3.172.233:9999
2014-10-11 04:27:54 send version message: version 70040, blocks=0, us=110.32.196.168:9999, them=192.3.172.233:9999, peer=65
2014-10-11 04:27:54 partner 192.3.172.233:9999 using obsolete version 70038; disconnecting
2014-10-11 04:27:54 ProcessMessage(version, 104 bytes) FAILED
2014-10-11 04:27:54 disconnecting node 192.3.172.233:9999
2014-10-11 04:27:54 trying connection 54.88.23.87:9999 lastseen=5.0days
2014-10-11 04:27:59 connection timeout
2014-10-11 04:28:00 trying connection 54.68.0.111:9999 lastseen=4.6days
 
Here are what I believe would be the steps to reproduce this...

1. Have active masternode on mainnet.
2. Open port 19999 (*do not close 9999)
3. Add testnet=1 to the conf.
*4. Put a "#" in front of the line with the key in the conf.
5. Restart darkcoind before 30 min memory on mainnet has expired.
6. Continue to monitor on mainnet to see if the network rejects the node.

* I think you have to put an "#" in front of the line with the key when you go to testnet or else darkcoind will not start.
 
Last edited by a moderator:
Think about it this way.. the keys in the conf and vin are used to verify that the masternode is proper, but only when the masternode is initially activated. Then the network checks to see that the IP is still online with open port to continue regarding it as valid. Perhaps there is still some communication with it after that point, but it appears that it being on testnet does not cause it to fail that check.

* The vin would obviously be continually checked, but switching to testnet would not affect the vin.
 
Last edited by a moderator:
It's looping through this:

2014-10-11 04:27:53 Misbehaving: 54.189.202.6:9999 (0 -> 1)
2014-10-11 04:27:53 ProcessMessage(dseep, 116 bytes) FAILED
2014-10-11 04:27:53 partner 54.189.202.6:9999 using obsolete version 70038; disconnecting
2014-10-11 04:27:53 ProcessMessage(version, 104 bytes) FAILED
2014-10-11 04:27:53 disconnecting node 54.189.202.6:9999
2014-10-11 04:27:54 trying connection 192.3.172.233:9999 lastseen=5.7days
2014-10-11 04:27:54 connected 192.3.172.233:9999
2014-10-11 04:27:54 send version message: version 70040, blocks=0, us=110.32.196.168:9999, them=192.3.172.233:9999, peer=65
2014-10-11 04:27:54 partner 192.3.172.233:9999 using obsolete version 70038; disconnecting
2014-10-11 04:27:54 ProcessMessage(version, 104 bytes) FAILED
2014-10-11 04:27:54 disconnecting node 192.3.172.233:9999
2014-10-11 04:27:54 trying connection 54.88.23.87:9999 lastseen=5.0days
2014-10-11 04:27:59 connection timeout
2014-10-11 04:28:00 trying connection 54.68.0.111:9999 lastseen=4.6days
Yeah, i need to update the dnsseeder^^
 
Anyone with a MN on mainnet have 10 min to check this out??
https://darkcointalk.org/threads/v0-15-testing.2611/page-9#post-24738

I'm talking about a vulnerability that currently exists for the integrity of darksend selection on mainnet people.
Have not checked yet, but as soon as mainnet masternode stops answering to respond to darksend ping requests (closed port and/or stopped darkcoind) , the network should remove it from the masternode list after 30mins - regardless of a running testnet node.
 
Have not checked yet, but as soon as mainnet masternode stops answering to respond to darksend ping requests (closed port and/or stopped darkcoind)
Leave port 9999 open and darkcoind is running because it is on testnet.

* Just trying to help here, and telling you, I am receiving payments on mainnet. I don't mean to sound flippant.
 
Updated testnet dnsseeder to protocol version 70040 - blockchain sync should work more smoothly again now :)
 
Leave port 9999 open and darkcoind is running because it is on testnet.
Good point! You could run a testnet node with 'rpcport=9999' to mimic a valid mainnet node, not sure if mainnet network would get that. Nevertheless you'll need a valid 1000DRK vin to initally start off the mainnet node. So for a successful attack you would need 100x 1000DRK to start at least 100 mainnet masternodes.
 
Good point! You could run a testnet node with 'rpcport=9999' to mimic a valid mainnet node, not sure if mainnet network would get that. Nevertheless you'll need a valid 1000DRK vin to initally start off the mainnet node. So for a successful attack you would need 100x 1000DRK to start at least 100 mainnet masternodes.
Exactly. It only compromises the speed of darksend mixing.

*Unless someone figured out a more useful way to exploit.
 
Exactly. It only compromises the speed of darksend mixing.
Yeah, you are right, it's a bug - testnet node should not be able to keep a mainnet list entry alive :)
Is this http://jira.darkcoin.qa/browse/DRK-99 ?

darksendping for mainnet should be invalid for testnet and vice versa to avoid that (Just like testnet addresses are invalid for use in mainnet due to different address version)
 
Status
Not open for further replies.
Back
Top