Hmm… At least in the last two logs, I don’t have any of those k-bucket messages.
Oh, but do you have different log level, as I don’t have any of TRACE either?
My node died about 2am so I just started four nodes and all have stalled after a few initial messages including three errors in the logs.
EDIT: After two hours with no log activity I’ve killed all my nodes.
6 Likes
I tried downloading that. I got:
Downloading file "testnet.exe" with address 332f45ae9482a8e03b5d5579f15fd9c8806fe4a2b30da2bc6e5ebb9a98fe4095
Did not get file "testnet.exe" from the network! Network Error Provider record was not found locally.
But a later file succeeded:
safe --peer=/ip4/178.128.166.245/udp/36572/quic-v1/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak files download -- "safenode.exe" b0bd3fcef1afbf917a9b8b2936c13c9f66f6243d542643cb4353644572b5e8bc
Starting logging to directory: "/tmp/safe-client.log"
Instantiating a SAFE client...
Downloading file "safenode.exe" with address b0bd3fcef1afbf917a9b8b2936c13c9f66f6243d542643cb4353644572b5e8bc
Successfully got file safenode.exe!
Writing 38051762 bytes to "/home/topi/.safe/client/safenode.exe"
2 Likes
Do anyone making downloads and uploads now?
I see large consumption of CPU (10%), network (~1MiB/s) and RAM (2.7GiB) resources and wonder if that happens because of user activity or it’s more like background noise.
upd. I stopped node after it hit 3GiB mark. I saw no leak previously, but I see it now.
If someone have RAM chart for 2 days of uptime, it would be nice to compare with what I observed.
1 Like
I’m getting
[2023-05-06T16:08:41.139005Z ERROR safenode::client::api] Failed to dial 12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak: DialError(Transport([("/ip4/178.128.166.245/udp/36572/quic-v1/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak", Other(Custom { kind: Other, error: HandshakeTimedOut }))]))
when running: ./safe --peer=/ip4/178.128.166.245/udp/36572/quic-v1/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak files download – testnet.exe 332f45ae9482a8e03b5d5579f15fd9c8806fe4a2b30da2bc6e5ebb9a98fe4095
When trying to use the download example from the OP. I assume that fail to dial means it’s unreachable? This is using WSL2 on win10.
Edit1:
Same thing using the windows binary straight through windows.
Edit2:
Also getting the same thing from my VPS.
Edit 3:
I realized I had my VPS behind a NAT, I moved it to full static IP with no firewall, nothing blocking it. Still the same result.
As bad hardware is one variable that could cause problem to testnet nodes, the thought was that maybe devs who investigate example a node failure, that it would help them by knowing that it is not hardware realated. But maybe it does no matter.
And it could also maybe help testers knowing why their nodes fail.
2 Likes
I had 9 nodes up. One died overnight, the rest are all up and seem to be doing fine, so far as I can tell.
4 Likes
It would if this was a likely failure mode but it isn’t. Memory corruption due to faulty hardware is rare.
1 Like
I’m not in a position to participate, but it looks fabulous. Absolutely thrilled to see the action here. Well done all
5 Likes
I also tried to upload the entire folder where the files were located but it didn’t download either.
Today I try to join the joshnet without success, tomorrow I will try again with more calm
3 Likes
I can see a value in a combined healthcheck and sizing script to check and report on things like the architecture, dependencies and also recommend how many nodes could be run on the computer when the CPU and RAM requirements are known.
I’m not sure it’s necessary/worth it to check the health of the RAM though. DIMMs going bad are rare these days and if there are RAM problems I think it will manifest in the system randomly crashing before you even get to install this software.
5 Likes
I used bad DDR 4 RAM sticks for 6 years, only noticed some problems when did SFC scan now and it often complained about corrupt windows files, so it seems it can go quite unnoticed for a long time.
SFC and DISM are often finding errors in windows cause windows is at fault not the RAM! a bad RAM would immediatly show its teeth by crashing apps or the system
1 Like
I am down to 4 running nodes, the rest no longer had log activity so I put them down.
there was a payment issue with Hetzner and I can no longer use it for safe so I used DigitalOcean, let me just say the price difference is a lot. I put down the droplet so -8 nodes. lets implement NAT traversal so people can host nodes at home
3 Likes
Did you try from home? I could run a node from a hotel.
Just tried from airport wifi but got the HandshakeTimedOut error. Could just be the network is on the way out.
4 Likes
I was getting that early this morning and its still the same
4 Likes
when I went to 200mbps they gave me a router that is shitty, I cant port forward correctly DMZ doesnt work, and pnp also doesnt work!
2 Likes
My node lost connection to network at least about three hours ago, maybe earlier, but there are no logs before that. It also seems, that about an hour after that my laptop crashed / restarted because of some unknown reason.
I restarted my node just now and it stopped right away. Last message in the log is this:
[2023-05-07T11:03:12.764036Z ERROR safenode::node::api] Failed to dial 12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak: DialError(Transport([("/ip4/178.128.166.245/udp/36572/quic-v1/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak/p2p/12D3KooWPuaKwn7rFNn4oj1ABDpikVZ7scHs29ZNCHBveLHgCmak", Other(Custom { kind: Other, error: HandshakeTimedOut }))]))
2 Likes
That’s a shame to end up with bad equipment after a bandwidth increase. The issue I had was 30% packet loss even with wired Ethernet!
There may be a way out of it though like I used. Does the router they gave you have a modem mode? If so, you could switch it to that and get something decent to do the actual routing to their network and form the one at your end.
2 Likes