Re: How (and why) to use the Relay Network

[Paladin45]

I'm trying this with P2Pool and Armory, Linux Mint 17 64-bit. I'm seeing "Unable to resolve relay peer hostname" messages in the terminal window. Are these errors or normal operation?

[1714964195]

1714964195

[1714964195]

1714964195

[Matt Corallo]

I'm trying this with P2Pool and Armory, Linux Mint 17 64-bit. I'm seeing "Unable to resolve relay peer hostname" messages in the terminal window. Are these errors or normal operation?

Sounds like you mis-typed the relay node server (or your DNS is messed up)?

[Paladin45]

I'm trying this with P2Pool and Armory, Linux Mint 17 64-bit. I'm seeing "Unable to resolve relay peer hostname" messages in the terminal window. Are these errors or normal operation?

Sounds like you mis-typed the relay node server (or your DNS is messed up)?

Yep. I missed the "relay" in the address string. Checked it a couple of times before I posted, but missed both times. Thank you!

[spin]

Getting lots of

Code:

Closing relay socket, failed to read message header (0: )
Closing relay socket, failed to read message header (104: Connection reset by peer)

On the EU node.  us-east working fine for me.

[PatMan]

Getting lots of

Code:

Closing relay socket, failed to read message header (0: )
Closing relay socket, failed to read message header (104: Connection reset by peer)

On the EU node.  us-east working fine for me.

Yup, same here with the EU node.......

[naplam]

I decided to connect to two nodes and forget about it. One will usually be working...

[roshii]

Yup, same here with the EU node.......

same issue for two days now, I'll switch to us-east this evening

[kano]

Getting socket errors every so often on the us-west here.

(the main issue with this is that it just sits there stalled for quite a while until it spits out the socket error message)

[PatMan]

Some kind of auto-switching to a working/fastest/less loaded server would be invaluable too......?

[kano]

Getting socket errors every so often on the us-west here.

(the main issue with this is that it just sits there stalled for quite a while until it spits out the socket error message)

Seems to be all OK now - no errors in my screen scroll back buffer (which is about 20 minutes) but there were regular ones last night.
I have 2 nodes, current pool and where I'm moving it, and they are both now putting out messages in sync.

[kano]

Some sort of oversize error (repeatedly) going on at the moment.

Getting this each time:
"Closing relay socket, got freely relayed transaction too large (0: )"
after a whole lot of txns (including some > 25k in size)
(about every 20s)

A ddos?

Edit: 20 minutes later still happening.
The main issue with it is that it's passing all the spam to bitcoind that's then rejecting it all.
Anyone know what's going on?

Looks like I'm gonna have to stop my relay

Edit2:
Yep had to kill it:
2014-10-26 12:04:25 Misbehaving: 127.0.0.1:45659 (10200 -> 10300) BAN THRESHOLD EXCEEDED
2014-10-26 12:04:25 Warning: not banning local node 127.0.0.1:45659!

Since the relay is local, bitcoind wont ignore it.

Edit3:
Restarted the relay and the spammer has gone now.

[roshii]

us-east did not work for long and shows same issue as eu 's one since about this morning
us-west working correctly

Code:

Connected to bitcoind with version 70002
Finished connect handshake with bitcoind
Closing relay socket, failed to read message header (10054: Unknown error)

[roshii]

eu server is back online

[galenecrosby]

Many of you have likely seen the recent discussions on the P2Pool thread and the post on the foundation blog about the relay network.

It exists as a way for pool operators (and all miners, though not hashers) to get their blocks relayed quickly across a separate network both as a backup to the P2P network and as a quicker way to get the latest blocks as it skips relaying transactions which have already been seen. Thus, if you're a miner anywhere from a small p2pool miner to a large pool, you should be running one of the relay network clients.

You can see more information about its original goals and its original announcement here

Current connection information is as follows:

The relay network's domains are designed to be a bit aggressive to allow for future growth and are laid out by region. The regions are us-east, us-west, eu, sgp, au, jpy, and hk (ie North East US, West Coast US, Western Europe, Singapore, Australia, Japan and Hong Kong). Currently, several regions map to the same set of nodes (all of the asian ones are currently hosted in Singapore), but in the future hopefully things will be more distributed so that you can get nodes closest to your bitcoinds.

You should use the Java relay network client available at https://github.com/TheBlueMatt/RelayNode/raw/master/client/RelayNodeClient.jar or the C++ client at https://github.com/TheBlueMatt/RelayNode/tree/master/client. This will connect using a specialized protocol which will skip relaying transactions which you have already heard about, saving significant bandwidth (and time).

Note that a version of p2pool with a python client integrated directly is available at https://github.com/TheBlueMatt/p2pool

To connect to the relay nodes, you should run RelayNetworkClient.jar or relaynetworkclient with the parameters public.REGION.relay.mattcorallo.com followed by your local bitcoind's address.

eg
$ java -jar ./RelayNodeClient.jar public.us-east.relay.mattcorallo.com 127.0.0.1:8333
or
$ ./relaynetworkclient public.us-east.relay.mattcorallo.com 127.0.0.1 8333


If you are a large pool, you should go to https://docs.google.com/forms/d/1UL82QdcXXEhZwSHJAK04Sk_cWg4zLOu8a216nO7Mt8c/viewform to sign up to get a private token which will be used in case of DoS, allowing the relay network to split into private and public parts.

Further details:
 * The relay nodes do some data verification to prevent DoS, but in order to keep relay fast, they do not fully verify the data they are
relaying, thus YOU SHOULD NEVER mine a block building on top of a relayed block without fully checking it with your own bitcoin validator (as you would any other block relayed from the P2P network).
 * The relay nodes all relay among themselves very quickly, so there is no advantage to peering with more than one, in fact, the increased incoming bandwidth during block relay spikes may result in higher latency for your nodes. Please only connect to relay nodes in your region (with multiple peers, one per bitcoind, if you have bitcoinds in multiple regions).
 * The relay nodes are NOT designed to ensure that you never miss data, and may fail to relay some transactions. The relay nodes are NOT a replacement for having peers on the standard P2P network.

You can find the source for the relay nodes at https://github.com/TheBlueMatt/RelayNode

If you have any comments/concerns/suggestions, please do not hesitate to ping me on IRC (BlueMatt on freenode).

Im also running a relay node client and it need a huge bandwidth..thanks for sharing to us this amazing and infomative post matt!!

[jedimstr]

US East Coast relay seems to be down.  West is fine.

[amaclin]

US East Coast relay seems to be down.  West is fine.

Confirmed east=down/west=fine from my side

[Matt Corallo]

US East Coast relay seems to be down.  West is fine.

Ugh, yea, I think it ran out of fds, but I'm adding a bit more logging to be sure...It has been restarted and is working now, but I'll watch for further issues.

[Matt Corallo]

We now have an Australian node! Those of you already using the 'au' region should (at max) simply need to restart your client to connect to it, otherwise switch to the au region. Thanks to forum member norgan for providing the server.

[MissouriMiner]

EDIT - It's working.  Dumb newbie (rushing) mistake.   I forgot to open 8336 outbound on my firewall (iptables).

I have a couple of newbie questions.  (I've been running p2pool for a long time, but not your relay client.)

How do we verify the client is working?  
 - I already had p2pool running, and ran "./relaynetworkclient public.us-east.relay.mattcorallo.com 127.0.0.1 8333" in a separate screen session.
 - I see "Closing relay socket, failed to connect() (110: Connection timed out)" every so often, but this appears to be normal based on forum discussion.

Do I need to port forward 8336?  
If yes, can I specify a different port?

[norgan]

Hey Matt, Aussie node is playing up.