The number of connected miners is changing/dropping

[Aprrentice]

the number of miners is dropping with time. it is not constant. mine starts with all 16 than keeps on changing to 12, 13, 14, 11, 9.... even though on the miner's side it shows that it is still connected. any suggestions or assistance? thanks.

[djfinch]

Is keepalive on miners on/true? Maybe some of them are slow and not able to send share in given time (shareTargeTime).

[Aprrentice]

yes, it is true. it is not easy and confusing how to choose the value of difficulty and shareTargetTime. have 16 miners 14 averaging 245Hs 1 at 200Hs 1 at 420 Hs any suggestions recommendations for diff settings and target time? my settings are min diff 10000 and targetTime 10. thanks.

[djfinch]

10000 is way too high as starting diff for 16 miners. Pool sent you ~21000 diff job (as per your screenshot). Then, because you are not able to submit anything back, pool sent 10000 diff job (maybe pool's mindiff). I'm looking at 2 last lines... You can try to calculate required diff for each miner and set starting diff a bit lower than diff for slower one but XNP using vardiff also toward miners so you can put literally anything there and wait one-two minutes to rebalance. I have slower miners and set 100... Yes, miner will rapidly spam proxy after first contact but it's not an issue when if they are not connecting/reconnecting very often. Hope this will help.

[Aprrentice]

thanks. got it. guess now it is performing better.

• estimating the hashrate of the miners at 3700. so as per the wiki x10 = 37000. So pool difficulty port at supportxmr is set to 5555 which is 20000.
• set minDiff to 1000
• set XNR port diff to 2000, 4000
• target time 10 (guess should tweak this number). is there a "formula" for that?

[bobbieltd]

Less headache, target time is from 5 to 15 (it should be lower than pool’s target time —> a myth) Keepalive will periodically tell pool or proxy that miners are still active —> no disconnection from pool’s side (except miners disconnect by themselves) I still do not understand the diff myths in crypto world and why it can possibly affect hashrate.

[djfinch]

@Aprrentice :

• regarding port... it does not matter. It is just about starting diff. You will get correct diff after a while.
• 1000 looks better.
• you can ignore another ports and their diffs when you using just one of them
• target time... it does not matter either. XNP creating job with diff which miner can solve in given target time. If you double target time then diff is also * 2 as miner have * 2 more time to solve it... So basically nothing changes except for network traffic. Anyway, XNP doing share validation so very short time may cause bigger overhead.

@bobbieltd : Myths? If your miner receive job which cannot solve and nothing else? In case pools diff is locked or vardiff not working? I'm still getting exactly the same fixed diff no matter how many miners are connected or how powerful they are... One ARM miner used just for testing can submit 4 to 6 shares per day :)...

[2019-08-17 11:32:14] #018 new job from pool.ird.semipool.com:22206 diff 250012 algo cn-pico/trtl height 349071

[bobbieltd]

@djfinch Hehe, sorry to be a dick. Because semi's admin is incompetent and not as good as MoneroOcean :_) . Diff was broken but for a newbie pool owner, it takes a lot of time to understand/know why.

[Aprrentice]

one more thing.

• So I have added a GPU rig of 2.5KHs to XNP. when the RIG connects directly to the pool, the diff is around 60000-78000 but on XNP, the total difficulty is not going above 24000 with almost most of it assigned to the RIG (22000).
• what does the 100% mean? Workers: 19, Hashrate: 6672 pool.supportxmr.com: 6672 H/s or 100.00% (24890 diff)

[bobbieltd]

1. Snipa (or supportxmr’s co-admin) gives a favor to XNP by giving lower diff. Direct connection : hashrate x 30 = 2.5k x 30 or around 60k-75k diff. XNP connection : hashrate x 5 or hashrate x 10 (I do not know exactly current implementation) then lower diff for XNP
2. 100% means XNP mines only at one pool (100% all miners to supportxmr here) and not multiple pools. XNP can mine at multiple pools at the same time.

[Aprrentice]

maybe off topic.

• I read the proxy is optimized and is good for low hashrate miners. would it be "affected" if in addition to 18 miners ranging from 100hs to 500hs, I also point a GPU rig of 10Khs?

• I am also finding this erro in the logs config

Malformed message from 60.191.xxx.xxx Message: HEAD http://112.124.xxx.xxx:63435/ HTTP/1.1 and when doing IPlookup, found out both IP address are in China!

[bobbieltd]

@Aprrentice

1. It is good for both small miners and big miners (small miners have better benefits but no harm to big miners). One of advantages is that you can switch pools (failover for example easier)
2. Blah blah scanning IP logs —> normal, forget/ignore it. It comes from IP scanning (blah blah malformed, socket error, socket timeout...)