Upload failed: Unexpected EOF

Node Operators troubleshooting
1

Hi Alexey,

Would you mind kindly letting me know where I can access the Services Applet?

Thanks!
Ben

2

Hello @lookitsbenfb ,
Welcome to the forum!

Start (or Search) button in your Windows, then type “Services”

See also Services (Services) - Win32 apps | Microsoft Docs

3

Amazing, thanks so much for the quick response!!

I’ve been running my node for 48 hours now, and have had two errors looking like the below. Is this something I should be worried about/require action?

Additionally, is there any sort of alert system that can tell me if I’m encountering errors that need my attention? Or is the best method just to ctrl-f my log for the keyword “error”?

Thanks again!

Screenshot 2021-11-15 103105
Screenshot 2021-11-15 1031051020Ă—175 19.4 KB

4

This is the same error as a canceled upload or context canceled, depending on when in the code the storagenode got a cancel from the client.
The meaning - your node loose the race for that piece and someone’s node was faster.

5

Thanks very much for another quick response, really appreciate it!

Is there any sort of alert system that can tell me if I’m encountering errors that need my attention?

Or is the best method just to ctrl-f my log for the keyword “error”?

Are there any other keywords I should search my logs for?
Thanks!

6

There is a lot of solutions: Topics tagged monitoring
For the periodic port check you can use this guide:

For the remote access to your dashboard you can use either

or [Tech Preview] Multinode Dashboard Binaries

The satellites will alert you by email if your node is suspended or disqualified (to receive them you need to specify a real email in your node’s configuration).
The suspension is not applied before disqualification, it’s applied only to unknown audit errors or if online score is below 60%.
So in most cases it’s too late for such emails. Thus do not count on them, if you want an reactive alarm system.

7

Thanks for these! I’ve read through them but it seems that the dashboard only provides data after the error has been discovered by a satellite - surely I would want to be aware of an error and fix it before it gets picked up by the satellite so that I don’t get penalised?

Is there any way to achieve that please?

Thanks!

8

I think you may to do some scripting with expected behavior, if it ever needed.
If your hardware is working as expected and you did not delete customers’ data, the node can work until the disk will die.
In the last case if you do not have any proactive monitoring for the hardware you can start from scratch.

The easiest way - just check your dashboard from time to time. If you are not sure in your hardware - then it would need to have a more close monitoring with specialized tools.

By the way, you can participate in [WIP] Email alerts with Grafana and Prometheus

9

Thanks very much for all your help, it really is appreciated!

1 Like