1.5.1 rate-limiting chain failures?

[svirpridon]

After taking a single connection failure, it seems like 1.5.1 isn't reconnecting properly and/or is triggering rate-limiting failures. In my logs I saw...

[22:41:14][0x00007f13eea4e740][info] nhmp | Disconnected
[22:41:14][0x00007f13eea4e740][warning] nhmp | read: End of file
[22:41:24][0x00007f13ea51a700][info] nhmp | Disconnected
[22:41:24][0x00007f13ea51a700][warning] nhmp | handshake: Connection reset by peer
[22:41:34][0x00007f13eea4e740][info] nhmp | Disconnected
[22:41:34][0x00007f13eea4e740][warning] nhmp | connect: Connection refused

I bounced the excavator server and it happily came up first try when I ran my driver script:

=========================== www.nicehash.com =========================
           Excavator v1.5.1a_nvidia GPU Miner for NiceHash.
           Copyright (C) 2018 NiceHash. All rights reserved.
=========================== www.nicehash.com =========================

Build time: 2018-03-28 15:07:22+02:00
Build number: 1835019509
[22:43:05][0x00007f70c585e740][info] Log started
[22:43:05][0x00007f70c585e740][info] core | Found CUDA device: GeForce GTX 970
[22:43:06][0x00007f70c585e740][info] http | Listening on 127.0.0.1:3456
[22:43:06][0x00007f70c585e740][info] core | Initialized!
[22:43:08][0x00007f70c585e740][warning] nhmp | close: Operation canceled
[22:43:08][0x00007f70c132a700][info] Connected to nhmp.usa.nicehash.com:3200
[22:43:08][0x00007f7099b57700][info] wrkr0-0 | Algorithm: CUDA-equihash parameters: M=1
[22:43:08][0x00007f70c585e740][info] nhmp | Subscribed
[22:43:08][0x00007f70c132a700][info] nhmp | equihash | add success
[22:43:09][0x00007f70c585e740][info] net | equihash | New job '000000332dd5de72', diff=1024
[22:43:10][0x00007f70c132a700][info] core | Algorithm 'equihash' speed: 174.034695 H/s
[22:43:10][0x00007f70c585e740][info] core | Device #0-0 speed: 173.066221 H/s (equihash)

[t4nja]

Can you reproduce the issue? If yes, can you post the logs here (run excavator with -f 0 parameter).

[svirpridon]

I have not seen it fail since unfortunately. It is possible it was network issues on my side, or a deployment on yours to nhmp.usa. If I'm the only one who saw it, you can always close and I can re-open if I find it again. That said I'd have to run in -f 0 all the time and hope, since when I bounced excavator, it worked.

[t4nja]

No need for the logs then. As you said it's probably the network or stratum issue. I'll close this for now and if you notice the problem again just comment here.

[dmatthewstewart]

I get this failure every 24-48 hrs on Nicehash running an Nvidia rig. It requires me to reboot. Restarting Nicehash does nothing at best. At worst, the cards are no longer detected. Running 9x1060's and it started happening only recently. Sometimes it seems to trigger a hardlock on the system requiring a forced hard reset

[t4nja]

@dmatthewstewart Which version of NVIDIA driver do you have?