Extremely high percentage of UnreadableRemoteFailure

[DerEwige]

eclair version=0.10.0 commit=a63d2c2

I recently noticed that about 80-90% of all payment end in an UnreadableRemoteFailure. This value seems way too high.

I only noticed this because I have a lot less data in one of my custom database tables, than what I used to have.

I have no idea when this started, but It looks like, it transitioned from mostly RemoteFailureto UnreadableRemoteFailurewithin the last 12 months.

I’m not even sure if this is a eclair issue or if maybe the problem is interop and got introduced in a recent LND version.

This is an example of a log line:

2024-07-28 16:00:00,487 WARN  f.a.e.p.s.PaymentLifecycle PAY n:023631624e30ef7bcb2887e600da8e59608a093718bc40d35b7a57145a0f3db9af h:c6a10a1f425dbebce0ca67f3d51124fd8df4c8a64f7b6e2f844b48f630e97ab9 p:5fab4964 i:1ebd97b8 - cannot parse returned error 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 with sharedSecrets=List((fa17cec915db5456e953e502ce9e22f76959a382d5416427a1a7cb4009da3cd6,035adfeca8fc95de5e2e9dbf9be39ac577b510e7dad0ad9287117d01f282eb667a), (dbf09689f56a69a4b865d5ef6763c86344e66868bf03c2b773ff27bac22762be,02d96eadea3d780104449aca5c93461ce67c1564e2e1d73225fa67dd3b997a6018), (c527e1740e0e74be44223a11a9cd033bd568cd50004e456f614e2316f1f0d0cc,02a49e170b9b5c0d4f4b4908da6fd797186f04166a0182f37ea87fb8523c3646ca), (36a7ef4e70a9e181de39ff18bfc708b99f8cec1ea13d92324cca8d2d7c12be25,033d8656219478701227199cbd6f670335c8d408a92ae88b962c49d4dc0e83e025), (a7c3ff9db72fc13e50a78e9cf86539d6a7f3304ecdc4e7c94a13d4936fd648f0,023631624e30ef7bcb2887e600da8e59608a093718bc40d35b7a57145a0f3db9af)): couldn't parse error packet=ByteVector(292 bytes, 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) with sharedSecrets=List((fa17cec915db5456e953e502ce9e22f76959a382d5416427a1a7cb4009da3cd6,035adfeca8fc95de5e2e9dbf9be39ac577b510e7dad0ad9287117d01f282eb667a), (dbf09689f56a69a4b865d5ef6763c86344e66868bf03c2b773ff27bac22762be,02d96eadea3d780104449aca5c93461ce67c1564e2e1d73225fa67dd3b997a6018), (c527e1740e0e74be44223a11a9cd033bd568cd50004e456f614e2316f1f0d0cc,02a49e170b9b5c0d4f4b4908da6fd797186f04166a0182f37ea87fb8523c3646ca), (36a7ef4e70a9e181de39ff18bfc708b99f8cec1ea13d92324cca8d2d7c12be25,033d8656219478701227199cbd6f670335c8d408a92ae88b962c49d4dc0e83e025), (a7c3ff9db72fc13e50a78e9cf86539d6a7f3304ecdc4e7c94a13d4936fd648f0,023631624e30ef7bcb2887e600da8e59608a093718bc40d35b7a57145a0f3db9af))

I get thousends of these every hour. (see attached log file: eclair.current.zip)

[DerEwige]

Looks like an issue in my plugin.

I did not handle the UnreadableRemoteFailurein my plugin. This meant that those tries would not register as failure and those hops would not get punished in my own path finding.

I fixed that and the % is dropping fast now at about 10% and still falling