WARN quinn_udp: sendmsg error: Os { code: 10040

Issue Report

Environment

  • Operating System:
    Windows 11
  • Pulsar/Advanced CLI/Docker:
    Advanced CLI
  • Network:
    Gemini 3G

Problem

2023-10-31T19:29:42.718902Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "A message sent on a datagram socket was larger than the internal message buffer or some other network limit, or the buffer used to receive a datagram into was smaller than the datagram itself." }, Transmit: { destination: 199.175.98.77:30533, src_ip: None, enc: None, len: 1326, segment_size: None }
2023-10-31T19:31:25.210887Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "A message sent on a datagram socket was larger than the internal message buffer or some other network limit, or the buffer used to receive a datagram into was smaller than the datagram itself." }, Transmit: { destination: 199.175.98.16:30533, src_ip: None, enc: None, len: 1326, segment_size: None }
2023-10-31T19:32:23.579696Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::single_disk_farm::plotting: Plotting sector (2.35% complete) sector_index=11

I’m receiving this warning message every 1-2 minutes. Plotting seems to be OK. Is this a problem?

The farmer, node and my network are using the default settings. I’m running the same configuration on MacOS and I haven’t received this message.

1 Like

Do you happen to have VPN enabled? VPNs usually have lower limits for packet size and can result in similar messages.

If not I’m curious what MTU do you have on the network interface with Internet connection on the machine that prints these messages.

Nope, no VPNs enabled. I will check the MTU though and report back when I can access that info. Unfortunately the one ISP in my area doesn’t let me change the router settings on their base plan, so I need to set that up first.

Also, it’s probably worth noting that I have 3 machines plotting concurrently on the same network, each with their own node. The one Mac is running smoothly, the two on Windows both receive this message.

MTU is on your local interface on the machine, not on router, especially since Mac doesn’t have that issue.

Also I’m curious if your public IP is IPv4, IPv6 or both. Tunneling IPv4 through IPv6 might potentially have interesting side effects.

Thanks. I’ll go check that in about 30 mins. I believe it is IPv4, but not 100% sure. I’ll post that info as well.

It looks like a few others on Discord are getting the same message. I’m open to try tunneling, but if someone who’s more experienced wants to chime in here first please do.

Looks like my public IP is IPv4

2023-10-31T22:19:23.807399Z  INFO subspace_farmer::commands::farm::dsn: DSN listening on /ip4/

Originally MTU was 1500

So far adjusting it to 7kb hasn’t stopped the WARN message

Do not adjust it unless you want issues, 1500 is standard and expected.

Must be something Windows-specific again, cc @shamil

2 Likes

Facing this on my Farmer aswell.

  • Operating System:
    Windows 11
  • Pulsar/Advanced CLI/Docker:
    Advanced CLI - Powershell
  • Network:
    • Gemini 3G
    • No VPN
    • iPv4 Public IP, no iPv6 IP Detected
  • MTU Info Below
PS C:\Users\Justin> Get-NetIPInterface | where { $_.NlMtu -ne 0 } | select InterfaceAlias, NlMtu

InterfaceAlias                               NlMtu
--------------                               -----
vEthernet (Default Switch (Wi-Fi))            1500
vEthernet (Default Switch (Ethernet 2))       1500
vEthernet (Default Switch (Ethernet))         1500
Ethernet 2                                    1500
vEthernet (Default Switch)                    1500
Local Area Connection* 2                      1500
Bluetooth Network Connection                  1500
Local Area Connection* 1                      1500
Ethernet                                      1500
Wi-Fi                                         1500
Loopback Pseudo-Interface 1             4294967295
vEthernet (Default Switch (Wi-Fi))            1500
vEthernet (Default Switch (Ethernet 2))       1500
vEthernet (Default Switch (Ethernet))         1500
Ethernet 2                                    1500
vEthernet (Default Switch)                    1500
Local Area Connection* 2                      1500
Bluetooth Network Connection                  1500
Local Area Connection* 1                      1500
Ethernet                                      1500
Wi-Fi                                         1500
Loopback Pseudo-Interface 1             4294967295

Let me know if you need help re-creating or testing, happy to use my PC as a testbed @shamil

1 Like

I decided to add a Windows farm to my setup - the node is on Ubuntu - and I am getting this same warning on Windows 10. VPN turned off. About once a minute.

The error could be connected to the “MTU path discovery issue with QUIC on Windows”. Here is the tracking PR in the upstream repository: feat(quic): enable path mtu discovery by mxinden · Pull Request #4750 · libp2p/rust-libp2p · GitHub

Also, QUIC for golang had a similar error a couple of years ago: disable Path MTU Discovery on Windows by marten-seemann · Pull Request #3276 · quic-go/quic-go · GitHub (cc @nazar-pc )

@skylordafk @ImmaZoni How often do you encounter the error?

1 Like

I have such error every 1 minute.

Do please let us know ASAP if the issue is likely to invalidate the plotting we’re doing while receiving those messages, even if a solution isn’t going to be immediately available, so we don’t waste our electricity on it.

I’m getting this too but not as often as everyone else for some reason. Win 10. No VPN. Advanced CLI.

It doesn’t have impact on your plot, it can’t invalidate it

Cool, thanks for confirmation.

Hey Shamil, I’m receiving this warning message every 1-2 minutes. Plotting seems to be OK.

For me im getting it very regularly, almost once every 30 seconds

2023-11-01T21:50:50.590221Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "A message sent on a datagram socket was larger than the internal message buffer or some other network limit, or the buffer used to receive a datagram into was smaller than the datagram itself." }, Transmit: { destination: 199.175.98.71:30533, src_ip: None, enc: None, len: 1326, segment_size: None }
2023-11-01T21:51:20.547689Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::reward_signing: Successfully signed reward hash 0x9756a28157c1d44660629b03445caae78e6ab78b2f1f842a301703056706d3c0
2023-11-01T21:51:51.366311Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "A message sent on a datagram socket was larger than the internal message buffer or some other network limit, or the buffer used to receive a datagram into was smaller than the datagram itself." }, Transmit: { destination: 199.175.98.12:30533, src_ip: None, enc: None, len: 1326, segment_size: None }
2023-11-01T21:54:31.608551Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::reward_signing: Successfully signed reward hash 0xb6da26c66cd9c945a997f8e75d5c7134089a9d7d62fdb79c9928a52adfd20ff6
2023-11-01T21:54:51.202266Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "A message sent on a datagram socket was larger than the internal message buffer or some other network limit, or the buffer used to receive a datagram into was smaller than the datagram itself." }, Transmit: { destination: 199.175.98.13:30533, src_ip: None, enc: None, len: 1326, segment_size: None }

Additionally, the reward hashes my farmer is signing are not present on the network as far as i can tell from RPC calls.

eg: 0xb6da26c66cd9c945a997f8e75d5c7134089a9d7d62fdb79c9928a52adfd20ff6 is not present on main chain so I might be on a fork.

I suspect that the reason the reward hashes aren’t showing is because incentives for 3g haven’t been turned on yet. Scheduled for tomorrow sometime I believe.

Windows 11 Pro
Advanced CLI
3g
No VPN

2023-11-02T20:57:07.241355Z  WARN quinn_udp: sendmsg error: Os { code: 10040, kind: Uncategorized, message: "Сообщение, отправленное на сокет датаграмм, было больше, чем буфер внутренних сообщений или был превышен иной сетевой параметр. Также возможно, что буфер для принятия сообщения был меньше, чем размер сообщения." }, Transmit: { destination: 199.175.98.74:30533, src_ip: None, enc: None, len: 1326, segment_size: None }
1 Like

Getting the message a lot less frequently now for no apparent reason. Where it used to be once every minute or two, I can now actually plot a few sectors between the messages.

For the record, the warnings didn’t seem to hurt farming any either. Signing reward hashes there consistent with what I’ve been getting on my Ubuntu boxes. That was true even when I was getting them every minute or two.