Testnet p2p - no transaction count

Staking & Delegation Operate a Stake Pool
21

All those files are generated before you register on the blockchain. So I have all of them and it runs in “core” mode, i.e. as a BP. The pool is registered on the blockchain now, but that doesn’t change the behaviour.

Either way, it can also run as a relay without those files, but I still don’t get p2p transactions.

As I mentioned before, when I turn off p2p and just run it as a non-p2p node, using the topology updater, then I get a transaction count. So there is something about p2p that changes the way the node runs. That’s my issue.

22

I thought you lost your keys…

In any case p2p works, it even works with exact config file what I posted, here. I tested by myself, not sure what to add more.

By the way, you registered your relay as 0.tcp.au.ngrok.io:16960:

And if I try to connect to it:

image
image1148×48 2.5 KB

So that is first thing what you should look into.

Without incoming connections you will not get cardano_node_metrics_txsInMempool_int metrics

23

Hi,

yeah, my port moves around each time I restart as I’m just using a free ngrok connection at the moment. My paid one is all used up for my mainnet pool. You can try connect to 19552. That’s the port it is running on now.

Do you have a non-p2p testnet node you can add that to the topology of? (0.tcp.au.ngrok.io:19552) It might be an interesting test to see if non-p2p still connect in to my node.

But with the p2p architecture I can change the port it refers to in the topology and I expect that this is the port that is reported to the network. I do receive incoming connections via p2p.

With non-p2p I can also change my port and update that via topology updater and get incoming connections, and transactions.

I know p2p works for others. But there is something in my setup, quite possibly the CGnat bypass that is causing an issue for the p2p. That is what concerns me.

24

You need incoming connection to work so other relays will connect to your BP/relay. Other relays will use port and hostname from blockchain in your case 0.tcp.au.ngrok.io:16960. So if 0.tcp.au.ngrok.io:16960 will not be accessible, that means you will not get incoming traffic and no cardano_node_metrics_txsInMempool_int will be there. I haven’t used topologyupdater for almost half year, but if I recall correctly you can push via it your IP and port and that why it works on non-p2p for you.

So bottomline - make your port available and everything will work.

P.S. Meanwhile I will add 0.tcp.au.ngrok.io:19552 to my relays, so shortly you should get incoming traffic from my 2 relays and then you will get cardano_node_metrics_txsInMempool_int

P.P.S
I will connect to your relay from p2p relay, it doesn’t matter if source is p2p relay or not it will work in anycase.

1 Like
25

Thanks, I realise about the need for incoming connections. That’s been my point, I do have incoming connections…but no transactions… see my earlier screen shot.

But do let me know when you’ve added the relay manually (p2p or otherwise). I think that will be a good test

26

Not sure what kind of incoming connection you have… If every time you restart your node port changes and topology is not brodcasting it, that means nobody is aware about your new port.

All relays from blockchain are trying to connect towards 0.tcp.au.ngrok.io:16960

1 Like
27

I can see I have new connections coming in now and the transactions are increasing…so I assume you’re connected to my node. Thanks

I guess the “5” incoming connections that I always get are not active nodes, or something like that. Or maybe some internal connections.

How many nodes have you connected? I can see 9 incoming now, i.e. 4 more than a few minutes ago. It will be interesting to see if the p2p starts announcing my node via yours.

28

Now it is just 1 Relay, I’m in process adding second one.

I put your relay in localRoots, so it will not be advertise. It is quite interesting though. Maybe I will put you later in publicRoots and will start to advertise your relay. It will interesting to see if you get traffic from other relays…

If you are checking connection with gliveview, maybe it do not shows it correctly. Try to check it with netstat.

1 Like
29

With my ngrok connection I can’t netstat the connections, as they all come in over the tcp tunnel, and then just get reported as a “local” connection.

But with your connection added and the transaction count going up, I’m happy that it isn’t an issue with the node or gLiveView. It is “only” a connection issue.

30

I’m up to 13 incoming now. So something is happening!!

31

By the way as temporary solution you can run your nodes in p2p and still use topology script just to push out your port and IP. So you will get incoming connections from pools what are using topology script and your node will connect to relays from blockchain using p2p features.

When all pools will switch to p2p from topology, that will stop work, though.

32

Hmm, I never thought of using the topology script (not updater) and p2p…interesting.

33

Hello there. I’m also facing a similar issue. I’m running PreProduction network pool to test my infra and other configs. Our ticker is tTurk (pool13ls4vfl0cznxjh0jm7kmzv5a5wf8e8uh27esucwm2f4mwt40sfv). I registered my pool with 2 Relay 1 BP. I download official configs from here I also turn on all of the debugging traces in config.json.

I have 50 outgoing connections on both relays (reporting in Prometheus via cardano_node_metrics_connectionManager_outgoingConns); however only 2 incoming connections, one of them is the other relay and the second one is the bp node. As a consequence, no TX processed and empty mempool, just syncing the chain. My log example is shown below.

ConnectionManagerCounters {
    fullDuplexConns = 0, 
    duplexConns = 49, 
    unidirectionalConns = 2, 
    inboundConns = 2, 
    outboundConns = 50
}

I don’t think that the issue is a firewall or connection issue. Am I missing something on configurations? I can share my config details as you wish. Thanks in advance.

PS: If anybody can add my relays manually, we can make the fastest test.

            {                                                                   
              "address": "turk-preprod1.cardanoturkiye.org",                                    
              "port": 2095                                                      
            },
            {                                                                   
              "address": "turk-preprod2.cardanoturkiye.org",                                    
              "port": 2095                                                      
            }
34

You could try the above, i.e. run the topology script, which notifies the other nodes that you exist. Once you get a few connections coming in that way, p2p should take over and get you more connections. Then you should start seeing transactions coming through.

35

I checked you relays and they are available from outside, so you should not have any problems with connections.

I see that you have new pool and if you have p2p enabled for me it took at least 1 week, before I started to get incoming connections, so you need to wait I would guess…

1 Like
36

Hello @os11k and @jeremyisme, thanks for your replies. First, I thought that it could be a misconfiguration or something else that I don’t know. So I didn’t try the topology updater script. I’ll do it.

Second, yes, I registered pool just a couple of days ago. It’s so strange that it takes 1 week to get connections. By the way, thank you for checking my relays.

If there is a chance that anybody can add my relays to their topology file manually, I’ll be very appreciated.

37

Hi, I add the topology updater script to my relay. After a couple of hours, I got a few incoming connections and processed TXs. However, I think they’re not P2P connections. I’ll wait a couple of days more. Thanks to everyone.

1 Like
38

Yeah, that’s how I got more p2p connections coming in. Once I got a couple of manually added connections coming in I got discovered on p2p easier and everything picked up.

39

Hi again. I’m waiting for 4 days, but no change in my incoming connection number. I got only one connection when I check via cardano_node_metrics_connectionManager_incomingConns. As Epoch 50 starts, I have to mint 3 blocks since it’s started. However, I missed all of them. is there anything I don’t know about the preproduction network?

07 Feb Update: I don’t know how it happened, however, my both relays have a lot of incoming connections, and process TXs and my BP node is minting blocks. I just missed the first 10 blocks :slight_smile:

1 Like
40

Hello Jeremy, sorry if i write in a closed ticket, but i think to have a similar issue

Total Tx: 0

i’ve already tested the P2P in testnet and everything was working fine, now i’m in mainnet, but i’m not having incoming peers and also transactions.
I’ve read this post, but i remember to have already tried to start the P2P and the topology updater in the same time and my node (when i was in testnet) was not working because both (topology updater and the P2P) seems in conflict.
Do you have some suggestions on how can i proceed?