Since November 14 build, sometimes get many thousands of "get_providers" errors when control-c'ing a farmer

[Qwinn1]

It seems to be linked to the piece cache sync issues since the 14th.

Example - I'll get thousands of these, totally explode the logs:

2023-11-17T12:29:19.926408Z WARN subspace_networking::utils::piece_provider: get_providers returned an error piece_index=4510 key=Multihash { code: 11770128, size: 8, digest: [158, 17, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0] } err=SendCommand(SendError { kind: Disconnected })

The attached log is for me trying to create a new farm on the November 16th build. Right now, it is pretty much impossible to sync cache on a new farm. This is on Ubuntu 20.04 on a 10850k, and I had even worse results on a better processor (7950x) in Windows (never even got a 2% complete there).

badcachesyncnov16.log

I have gotten the get_providers errors when control-c'ing established farms, btw, which tend to cause "missing piece" errors until the problem sometimes sorts itself out with repeated restarts.

Please note that the log only ended where it did because I was forced to kill -9 the farm to get it to stop spitting these out. On a couple of rare occasions, I only got a handful of the get_providers errors and the farm did eventually stop, but I don't recall it producing any more useful messages afterwards.