Crash when stratum connection lost - neoscrypt

[david-987]

Only happens with this fork of ccminer.. tried 8.15 through 8.18, no luck. Win 10, both connected to bsod.pw pool

[nollijimayu]

me, too

[KlausT]

I wish ccminer would crash here on my system, so I can fix it, but it's running just fine here.

[iamchw]

me,too

[dementeb]

Can confirm, crashes on lost connection (latest message - "Resetting GPUs") on my 3 rigs with Win10x64 388.71 driver (and with 390.65 to). Doesn't crash on test system with single 1080ti on Win2012R2 390.65

8.19pre both cuda version
-D doest show more info (
[2018-01-13 02:35:38] thread 2: continue with old work [2018-01-13 02:35:38] GPU #2: start=ab4caaaa end=ac4c3ee3 range=00ff943a [2018-01-13 02:35:39] accepted: 3/3 (100.00%), 1825.29 kH/s yay!!! [2018-01-13 02:35:42] stratum_recv_line failed [2018-01-13 02:35:42] Stratum connection interrupted [2018-01-13 02:35:42] restart_threads [2018-01-13 02:35:42] GPU #2: GeForce GTX 1060 3GB, 665.93 kH/s [2018-01-13 02:35:42] GPU #1: GeForce GTX 1060 6GB, 693.85 kH/s [2018-01-13 02:35:42] GPU #0: GeForce GTX 1060 6GB, 459.79 kH/s [2018-01-13 02:35:42] Stratum session id: deadbeefcafebabe628d000000000000 [2018-01-13 02:35:42] Stratum authentication failed [2018-01-13 02:35:42] ...terminating workio thread [2018-01-13 02:35:42] workio thread dead, exiting. [2018-01-13 02:35:42] stopping 3 threads [2018-01-13 02:35:42] resetting GPUs

[KlausT]

I have found a possible reason why it crashes.
If you would like to test it:
ccminer-819crashfix-cuda91-x64.zip

[Hawxxi]

Just made an Account to say im gonna test it now, i can tell you in about 2hrs if this is more stable
EDIT: works fine for me

[minergo123]

hello got same problem
with test fix file
win 10
gpu 6x1060
driver 388
ccminer 8.15

let you know in like 12 hr

[dementeb]

Unfortunate have same crash + new crash - now it's do "resetting GPUs" without work, just after attempt stratum connection - aut error - resetting GPUs - crash (I'l use simple looped .bat for mph profit switching).
Previous version crash only if worked before (after success stratum connect, work, and stratum error).

And its fully stable for me if no stratum error - all version. 24h-48h no problem.

[minergo123]

tested
crashed
the same error

[david-987]

@KlausT Tested 8.19 crashfix, same crash.

"[2018-01-15 10:57:14] Stratum connection interrupted
[2018-01-15 10:57:14] GPU # 2: GeForce GTX 1070, 1076.20 kH/s
[2018-01-15 10:57:14] GPU # 5: Gigabyte GTX 1070, 1110.98 kH/s
[2018-01-15 10:57:17] GPU # 4: waiting for data
[2018-01-15 10:57:17] GPU # 1: waiting for data
[2018-01-15 10:57:17] submit_upstream_work stratum_send_line failed
[2018-01-15 10:57:17] ...retry after 10 seconds
[2018-01-15 10:57:17] GPU # 0: waiting for data
lots of waiting for data lines
[2018-01-15 10:57:30] submit_upstream_work stratum_send_line failed
[2018-01-15 10:57:30] ...retry after 10 seconds
[2018-01-15 10:57:32] GPU # 4: waiting for data
lots of waiting for data lines
[2018-01-15 10:57:35] GPU # 5: waiting for data
[2018-01-15 10:57:35] Stratum connect failed: Failed to connect to pool.bsod.pw port 1997: Timed out
[2018-01-15 10:57:35] ...retry after 10 seconds
[2018-01-15 10:57:38] GPU # 4: waiting for data
[2018-01-15 10:57:38] GPU # 1: waiting for data
[2018-01-15 10:57:38] GPU # 0: waiting for data
[2018-01-15 10:57:38] GPU # 3: waiting for data
[2018-01-15 10:57:38] GPU # 2: waiting for data
[2018-01-15 10:57:38] GPU # 5: waiting for data
[2018-01-15 10:57:40] submit_upstream_work stratum_send_line failed
[2018-01-15 10:57:40] ...retry after 10 seconds
[2018-01-15 10:57:41] GPU # 4: waiting for data
lots of waiting for data lines
[2018-01-15 10:57:47] GPU # 1: waiting for data"

[KlausT]

Are you using the -r option?

[iplayminer]

i got same issue , keep crash when stratum connection lost connecting to uniminingpool

[david-987]

@KlausT I'm not, should I?

[david-987]

Wouldn't work with -r option, just hangs on 'waiting for data'. Btw you can force a similar crash by doing ctrl + c in the cmd prompt, it loses stratum and ccminer stops responding, might help with debugging.

[KlausT]

I was just wondering if ccminer crashes while trying to exit, or if the crash happens while waiting.
For some reason it doesn't crash on my system when I do ctrl-c

[david-987]

If you point to bsod now you should be able to see it, the pool is having connection issues. Just crashed again, it was quick this time, just 11 seconds after connection lost

[2018-01-15 11:49:30] Stratum connection interrupted
[2018-01-15 11:49:41] stratum_subscribe timed out
[2018-01-15 11:51:41] GPU # 1: waiting for data
[2018-01-15 11:51:41] GPU # 4: waiting for data
[2018-01-15 11:51:41] GPU # 2: waiting for data
[2018-01-15 11:51:41] GPU # 3: waiting for data
[2018-01-15 11:51:41] GPU # 0: waiting for data
crash

[iplayminer]

it crash while waiting

[david-987]

Just saw it successfully reconnect after losing connection on one of my rigs, so it doesn't happen all the time.

[KlausT]

This is really wierd.
Try this:
ccminer-819nexttry-x64.zip
If this doesn't help then I'm out of ideas what could be wrong

[minergo123]

Ty KlausT
i will test it to - hope its work

[georgeous30]

Just some info, To people who still crashing with neoscrypt, use program called "Restart on crash" while waiting mr. @KlausT to fix this problem.

[david-987]

@KlausT still crashing same way as previous version

[dementeb]

new fix does't fix on my system
ccminer-klaust-c91 -r 0 -i 16 -a neoscrypt - options in my case
but if I just close command window with working miner - all fine, after I can start zm or tpruvot or KlausT miner without problem - so, maybe don't needed "terminate threads, resettings GPUs" on lost connections?

[aalyokhin]

Can confirm, having the same issue unfortunately. 1070, CUDA 9, latest drivers, Win10x64.

@georgeous30 that app doesn't work for me when ccminer is crashing in this way, it just keeps recognizing it as 'running'.

[madcorp206]

Same issue here.

Infos:
Linux 4.13.0-21-generic #24-Ubuntu
ccminer 8.19-KlausT (64bit) for nVidia GPUs
Compiled with GCC 6.4 using Nvidia CUDA Toolkit 9.1

Log:
[2018-01-16 07:53:55] Stratum connection interrupted
[2018-01-16 07:53:55] GPU #5: EVGA GTX 1070, 921.07 kH/s
[2018-01-16 07:53:55] GPU #3: EVGA GTX 1070, 1020.58 kH/s
[2018-01-16 07:53:55] GPU #4: EVGA GTX 1070, 1020.16 kH/s
[2018-01-16 07:53:55] GPU #2: ASUS GTX 1070, 1020.17 kH/s
[2018-01-16 07:53:55] GPU #1: ASUS GTX 1070, 1003.20 kH/s
[2018-01-16 07:53:55] GPU #0: ASUS GTX 1070, 1015.61 kH/s
[2018-01-16 07:53:58] GPU #5: waiting for data
[2018-01-16 07:53:58] GPU #3: waiting for data
[2018-01-16 07:53:58] GPU #2: waiting for data
[2018-01-16 07:53:58] GPU #4: waiting for data
[2018-01-16 07:53:58] GPU #1: waiting for data
[2018-01-16 07:53:58] GPU #0: waiting for data

It happens after 1 to 8 hours of mining.

[KiparisD]

@madcorp206 almost same rig 31070 + 31070Ti and same crashes, one solution that i found - it's multiply instances of miner, and you must find your scheme, i.e 3 instances of miner by 2 cards in each (-d 0,1 + -d 2,3 etc.), or it can be 2 instances for 3 cards each, or four instances 2+2+1+1 card, idk what the system of this or what causing it, but it works, even if it lost connection, my cards reconnect after that and continious mining without crash of miner

[minergo123]

@KiparisD can you share example bat command pls like for 6x1060

[madcorp206]

@KiparisD For now I'm trying another fork of ccminer, no crash yet, but only 6 hours for now.

[alex-df]

Same issue here, running on 5x1070 + 2x1070 Ti, Windows 10

[madcorp206]

Same issue with another ccminer fork but error message in inifnite loop:
GPU # : result for nonce $00000000 does not validate on CPU!

[kornface13]

@madcorp206 try lowering your memory clock speed by 25 or 50 in Afterburner. I couldn't run this for more than 30 seconds, and this seems to have solved it. Going 10 minutes now without errors. We'll see if its still running in the morning.

[dementeb]

Subj not about O/C, its crash without any work on lost connection. If connection always present - its worked fine imho

[wesleychannon]

Same issue on Ubuntu 17.10 x64 using ccminer 8.19-KlausT with 4x 1070:

[2018-01-24 14:33:00] accepted: 4764/4771 (99.85%), 4243.50 kH/s (yay!!!)
[2018-01-24 14:33:09] accepted: 4765/4772 (99.85%), 4243.36 kH/s (yay!!!)
[2018-01-24 14:34:05] accepted: 4766/4773 (99.85%), 4238.29 kH/s (yay!!!)
[2018-01-24 14:36:05] Stratum connection timed out
[2018-01-24 14:36:05] Stratum connection interrupted
[2018-01-24 14:36:08] GPU #2: waiting for data
[2018-01-24 14:36:08] GPU #3: waiting for data
[2018-01-24 14:36:08] GPU #0: waiting for data
[2018-01-24 14:36:08] GPU #1: waiting for data
Segmentation fault (core dumped)

Seems specific to when the connection to the stratum server is lost. Was running fine for about 2.5 days prior to the crash.

[KlausT]

I don't know very much about debugging segfaults under Linux.
Maybe one of you could try the linuxdebug branch:
https://github.com/KlausT/ccminer/tree/linuxdebug

If this version is producing a segfault you could try something like
gdb ccminer core
or something like that, so that I know what in function it is crashing