Problems mining on ethpool - what's going on?

hasherhasher Member Posts: 642 ✭✭✭


One of my rigs is doing some strange things (as you can see above). Ethpool thinks its offline (no stats are appearing). It was working fine for a week until today. What is wrong? :(
«1

Comments

  • hasherhasher Member Posts: 642 ✭✭✭
    Is my DAG corrupted or something?
  • hasherhasher Member Posts: 642 ✭✭✭

    I have these files sitting in my DAG folder. I'm going to try and delete them all and see if it fixes anything.
  • hasherhasher Member Posts: 642 ✭✭✭
    edited March 2016
    FYI, I have 4 cards on this rig. 3 x 7990 and 1 x 280X. Been running flawlessly until now..

    This is my start up batch file:

    setx GPU_MAX_ALLOC_PERCENT 100
    setx GPU_USE_SYNC_OBJECTS 1
    setx GPU_MAX_HEAP_SIZE 100
    timeout /t 20
    qtminer.exe -s us1.ethpool.org:3333 -u [mining.address].[worker] -G --cl-local-work 256 --cl-global-work 16384 --cl-extragpu-mem 0
  • AetsenAetsen Member Posts: 30
    Having problems on other pools as well. Not sure what's going on. http://ethermine.org/ has a trojan virus in it's eth proxy zip file. ethpool.org may have one as well, but there is bigger issues going on with multiple pools not working.
  • workwork Member Posts: 2,084 ✭✭✭✭
    Pretty sure this is a bug with both genoil and qtminer's stratum implementations. I've seen it a few times. Doesn't seem to ever occur if you use eth-proxy. Restarting miner fixes issue.

    @Aetsen lol... it's a false positive, man. Eth-proxy has been used on a botnet or two, so AV companies started flagging it. Happens all the time. Stop spreading FUD about ethermine/ethpool.
  • AetsenAetsen Member Posts: 30
    @work Hasn't fixed it for me. I have tried ethminer, qtminer and eth-proxy getting errors on multiple rigs with all of them.
  • workwork Member Posts: 2,084 ✭✭✭✭
    edited March 2016
    @Aetsen What errors do you get using eth-proxy and web3-umbrela ethminer?
  • AetsenAetsen Member Posts: 30
    @work I am getting json errors -32700 and json database errors on everything and with multiple pools.
  • AetsenAetsen Member Posts: 30
    @work it has been working fine for over two weeks. Just started tonight.
  • workwork Member Posts: 2,084 ✭✭✭✭
    Sounds like stratum woes, servers giving incomplete responses. I've seen it happen with qtminer and genoil ethminer, but not with eth-proxy setup for multiple stratum connections.
  • AetsenAetsen Member Posts: 30
    @work Could be. Not sure what's going on. It will hash for a minute or two then starts returning errors and when it does hash it shows the server receiving 20% or so less hash than my rig is actually doing.
  • workwork Member Posts: 2,084 ✭✭✭✭
    @Aetsen can you post a log from eth-proxy where it's throwing errors? Make sure you have at least 1 failover setup.
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    @Aetsen @work eth-proxy here on a bunch of rigs - no problems.
  • workwork Member Posts: 2,084 ✭✭✭✭
    dlehenky said:

    @Aetsen @work eth-proxy here on a bunch of rigs - no problems.

    Ya, I've never seen this issue with eth-proxy. It reconnects after a stratum failure just fine from what I've seen. Qtminer and genoil's ethminer seem to enter a state where they think they are connected to stratum but actually aren't, and keep mining on the same work. Seems to happen when the pool provides bad json responses and the connection drops. I'm 90% sure it's a bug.
  • hasherhasher Member Posts: 642 ✭✭✭
    So I deleted the DAG. It started mining again for a few minutes, then the same thing happened again. The strange thing is that its not happening on any of my other rigs. If i'm not mistaken, @dr_pra runs ethpool/ethermine? Would be awesome if he could tune in on this issue :)
  • dr_pradr_pra Member Posts: 445 ✭✭✭
    We will try to deploy a fix for this issue today. In the meantime please switch to eth-proxy for an immediate workaround.
  • workwork Member Posts: 2,084 ✭✭✭✭
    edited March 2016
    @Hasher there was some talk about this in the ethpool thread already. @dr_pra commented that you might get kicked off stratum if your rig submits too many bad shares. But, there wasn't much other then that. This isn't just an ethpool problem. It happens on all pools I've tried recently, but some more often then others.

    The solution seems to be to go back to eth-proxy and use getwork from the miner.
  • hasherhasher Member Posts: 642 ✭✭✭
    @dr_pra thanks, much appreciated!
  • GenoilGenoil 0xeb9310b185455f863f526dab3d245809f6854b4dMember Posts: 769 ✭✭✭
    @work @dr_pra can any of you elaborate on the problem a bit more and what I could possibly do about it?
  • dr_pradr_pra Member Posts: 445 ✭✭✭
    @Genoil need to investigate in detail, most likely the client acts up on some malformed json package.
  • hasherhasher Member Posts: 642 ✭✭✭
    @dr_pra did you make any changes on your side of the pool server in the past 2 days? If you did, thats probably what messed it up, because I've never had this problem before. But as I said, only one of my rigs has acted up. All others are still mining perfectly fine.
  • taz002devtaz002dev 0x26ff3ed5c5025b5a3c288b9139d0396f9de5891fMember Posts: 47
    guys where from can I get ethminer.exe ? Not genoil's one. Just neeed to do plenty of testing because of latest problems with this pool. One of my rig is getting erros and I wanna know why.
    Thanks
  • BiodomBiodom Member Posts: 693 ✭✭✭
    @hasher I also start getting more errors than before on qtminer on two out of three rigs, which never happened before recent updates. Rigs are still hashing, but error accumulation is bewildering.
  • taz002devtaz002dev 0x26ff3ed5c5025b5a3c288b9139d0396f9de5891fMember Posts: 47
    edited March 2016
    Always used qtminer and I always got lots of errors on this pool while on others I don't.
    Eg. :coinotron...
    Mentioned that before in a number of comments and was told my rigs are heavily tuned while they are actually under tuned... Will step out a bit because don't like what I see
    Biodom said:

    @hasher I also start getting more errors than before on qtminer on two out of three rigs, which never happened before recent updates. Rigs are still hashing, but error accumulation is bewildering.

    Post edited by taz002dev on
  • hasherhasher Member Posts: 642 ✭✭✭
    We need to find a solution ASAP as I (and i'm sure many others here experiencing similar problems) are losing valuable ROI time on our equipment with all this downtime, and the increasing block difficulty isn't gonna wait for us...
  • workwork Member Posts: 2,084 ✭✭✭✭
    edited March 2016
    hasher said:

    We need to find a solution ASAP as I (and i'm sure many others here experiencing similar problems) are losing valuable ROI time on our equipment with all this downtime, and the increasing block difficulty isn't gonna wait for us...

    cc: @taz002dev

    You've been given a solution... switch to eth-proxy and web3-umbrella ethminer.

    https://github.com/Atrides/eth-proxy/releases
    https://github.com/ethereum/webthree-umbrella/releases

    Lost stratum connections don't seem to happen with eth-proxy, only in the miners with stratum built in. Plus, with eth-proxy you have stratum failover.

    I prefered QtMiner or Genoil's ethminer before, but with all pools I've tried dropping more and more connections lately, I've moved my rigs back to eth-proxy.

    @dlehenky do you get GPU usage drop-off for a fraction of a second everytime a share is found? That's my one annoyance with eth-proxy/ethminer. Not overly significant when share target is fairly high, but it's definitely lost hashing power compared to qtminer/genoil.
  • dr_pradr_pra Member Posts: 445 ✭✭✭
    The issue should now be resolved. Please report if you still experience lots of dropped connections on the pool.
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    @work I run eth-proxy per rig, with farm-recheck 50 on ethminer, and my kernel runtime at 40 ms (global work 16384), so no I don't notice that drop-off.
  • workwork Member Posts: 2,084 ✭✭✭✭
    edited March 2016
    @dlehenky I'm running 50ms farm-recheck too, and I get 2 periods on 0H/S everytime there is new work or a share is submitted. If farm-recheck is 200ms, I still get 2 periods of 0H/s, same bumping it to even 1000ms for testing - 2seconds of 0H/s. This is originally why I stopped using getwork to a stratum proxy, because I can get qtminer or genoil working with no drops.

    Edit: 256th post - my kind of number =]
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    @work And how long is your kernel runtime? I have any idea from your previous posts what it is, and you're running a much longer kernel than I am. I build my ethminer from source, and I have my own per block timer and kernel execution counter in the c++ code that calls the kernel; I don't rely on the time sampled hash rate numbers that ethminer produces. I know exactly how many kernels run on each GPU on each block and exactly how long they run.
Sign In or Register to comment.