ot-node icon indicating copy to clipboard operation
ot-node copied to clipboard

feat: add external ip to network config

Open zeroxbt opened this issue 11 months ago • 4 comments

I have tested on multiple community nodes, and it appears that 14 out of the 32 nodes in gnosis:100 sharding table are not reachable as their public IP is not known. I suspect that these are either AWS or GCP nodes, but this needs to be confirmed. The list of node ids is:

QmSZqi5YCEFVWSgCVd6id7dZor5bhxPcPg1B8KKRbaDt3w
QmdxdGWjRqtE2wdoaoWQFMqrUPrsnFoLJ3QcCFDJ3fPtTt
QmeaeLpQENK5eHHMF6Kt5iqeFtqADV8vUzcp3jYeSghCQy
QmQvzCuUqpfxUELQXoYyiQWKWwtSt8UY8K6du9f9P4St1h
QmZ7xMtgJ1vBey5ifvD6yFsxH4VmSQVwtKrX8zNuZ2uiKb
QmZSFcfQ9DesB4ZjkT3qJ3VyCYDLFRYbC5U1VyVPN47dwE
QmapSytEE1LaK9kM7PYDeit8qDMYph5LvyaGBxxEYViCr5
QmQYb2ZXw75gJNvqobpe8gaYphZ7BBuQHxGfFg7JtKra8Z
QmTgmcMUCpsJoontxtNuFapNCUNTVVQ4QG6JM4HZqbYsJj
QmckQkHQEyrTvFGaTe3TJNRPJRyeGTHFawK159BWRgUGbm
QmQJXftuDhD1XENTdVh97CZEpQB7zpPpMBSUWRUVgMY4mm
QmSzGaTUu9VT7WdJcyX93SHsUDqXpnAGFxWQ8jdZLcnrL5
QmPes37bZ6PKDPBmQNJC88dCAkCEVxEsK9HuHFyGkN5Rfm
QmS4gHcTycnrBPXStsTq7AwNAsT3wUSksXXTkrhr5WapQe

This PR gives the option to node runners to add their node's public IP to the config file, so that the node can announce their public multiaddresses to other nodes when identifying on the network.

zeroxbt avatar Mar 06 '24 02:03 zeroxbt

Hi @zeroxbt, i can confirm that all problematic node ids are nodes running on AWS or GSP. We will perform some tests for your fix before including it to release. Thanks for the PR.

djordjekovac avatar Mar 06 '24 15:03 djordjekovac

@zeroxbt did you test this change. After pushing fix on AWS and GSP nodes i was still getting error Unable to dial pear on the network.

djordjekovac avatar Mar 07 '24 09:03 djordjekovac

@zeroxbt did you test this change. After pushing fix on AWS and GSP nodes i was still getting error Unable to dial pear on the network.

I haven't tested this change as I don't have access to AWS or GCP nodes, and also don't know how these nodes were setup on your end. The idea was to let you add external IPs and see if that would help with connection issues. What I can do is include a command that periodically logs public addresses of peers from sharding table, so you can verify on devnet if there is any improvement.

zeroxbt avatar Mar 07 '24 10:03 zeroxbt

@djordjekovac I've added the command, which will only log on development and devnet environments. It would be good if you could review with @Niks988 how these nodes are setup, and ensure they are not behind NAT or double NAT.

zeroxbt avatar Mar 07 '24 10:03 zeroxbt

Hi @zeroxbt, we did another round of testing of your solution and we can now confirm that this fix is working. We applied this patch on all nodes that are behind NAT and status of the network improved a lot. I added one comment on the code so please address it before we merge. Thanks again for helping!

djordjekovac avatar Mar 29 '24 09:03 djordjekovac