View Single Post
  #547  
Old 31st December 2017, 10:01 AM
ainarssems ainarssems is offline
Senior Member
 
Join Date: Aug 2010
Location: Rushden, Northants
Posts: 3,847
Default

Quote:
Originally Posted by MikkiJayne View Post
Yeah it did look like speculation. If the price is sensible and we find out before my next Vegas materialise from Amazon.fr then I'd be up for a Giant B.

Who else is running Cryptonight by the way? I get a lot of 'SOCKET ERROR - RECEIVE error: socket closed' errors in both Cast-XMR and XMR-Stak (several versions of this one too including current), at which point it stops mining and waits to reconnect. It could be a minute or two before it restarts. This happens regardless of which pool its on, although Nicehash is much worse than MiningPoolHub, and it hardly did it at all on Electroneum so I wonder if its related to difficulty? Its all over the place on Github and Reddit but there are no solutions - just comments that if you're not finding shares fast enough the pool will disconnect, but it happens way more often than the alleged 10 minute timeout. 3.7KH/s really ought to be fast enough not to time out!


While I've been mining for myself I've just put up with it, but its a problem on MiningRigRentals because it seriously messes with my average hash rate thusly:



Each of these valleys is one of these events.

Ainars - I had a look at your rigs on MRR and they appear to exhibit the same behaviour. Have you noticed this error?

Has anyone else seen / solved it?
Yes I have noticed disconnects, it's mostly depending on pool on which the renters mine. Sometimes I get a lot of them sometimes none, I did notice that MRR Amsterdam server sometimes gives a lot of disconnects, switched to de server and problem went away.

The dips in the graph does not always correspond with disconnects as I am getting a lot of dips without disconnects as well. It's just the way pool calculates hashrate on submitted valid shares and not the actual work your rig is doing. That's why the peaks show higher hashrrate than the rig could possibly do. It all averages out about where it should be, I guess there is no hashrate reported from rig to the pool as that would be open to abuse where somebody could modify miner to report incorrect higher rate to the pool than it's actually doing so the pool has to rely on calculating the rate from the shares you submit and that changes over time depending on luck so sometimes pool reports higher, sometimes lower or sometimes none at all.

I do sometimes get renters with a crappy pool that mines with low static difficulty or gets a lot of outdated shares which don't count towards hashrate or whatever and it reports poor hashrate. In those cases I either point my single RX580 card that runs on Microserver and is normally on NH to that rig to give it a bit of boost or extend rental a bit so that it averages at least 90% to give positive impact to rating. Or sometimes if it looks very poorly I just send a message to renter saying this is not working, sort it, mine on different pool or whatever or I will terminate the rental. If they don't I send another message to MRR support to end the rental, they refund money for unused time to the renter, I keep the money for used time and I don't get negative impact to my rating even if reported hashrate was low.
__________________
Currently 8less
2011 Q7 S Line 3.0TDI, 2016 Tesla Model S 90D

8 history:
2006 A8 Sport 4.2TDI quattro SOLD,
1997 S8, reached end of life with gearbox failure

Last edited by ainarssems; 31st December 2017 at 10:30 AM.
Reply With Quote