Powerline will most likely not work because my router is in my flat. My miner is in my garage.
The cable goes from my flat to a distributor. From the distributor to the switch and from the switch to my garage. With several fuses between my flat and the garage.
It would be a miracle if powerline would work here.
Topic: ETH stale shares (Read 193 times)
<100 ms response time is where you want it. <20 ms is excellent. WiFi is notoriously unreliable if you live in a metro area with many access points connecting to the same default band channel, especially if you are connecting on a 2.4GHz band has a limited number of channels which can be very congested. You can use a network analyzer tool to see how congested the WiFI channel you are connecting to is and try connecting to a less congested channel in your access point.
https://play.google.com/store/apps/details?id=net.techet.netanalyzerlite.an
Connecting to a building outside of where your access point is located can also be very unreliable. The longer the distance from the access point and the more obstacles the the connection goes through, the more unreliable it is. You can look in to Powerline networking as an option, which is significantly more stable than WiFi.
https://www.digitaltrends.com/computing/everything-you-need-to-know-about-powerline-networking
https://play.google.com/store/apps/details?id=net.techet.netanalyzerlite.an
Connecting to a building outside of where your access point is located can also be very unreliable. The longer the distance from the access point and the more obstacles the the connection goes through, the more unreliable it is. You can look in to Powerline networking as an option, which is significantly more stable than WiFi.
https://www.digitaltrends.com/computing/everything-you-need-to-know-about-powerline-networking
Good point.
I just try an other pool. Let's see if I get the same problems here.
EDIT:
Withing 5 ms i get always the exact same ping from flexpool.
I just try an other pool. Let's see if I get the same problems here.
EDIT:
Withing 5 ms i get always the exact same ping from flexpool.
This must be a problem with ethermine...
Is there anything I can do? Except changing the pool
Is there anything I can do? Except changing the pool
I wouldn't make that assumption. Changing pools would prove if it's a pool problem.
You can't solve the problem until you know what it really is. Don't dismiss a wifi issue.
I mine on the EU pool from ethermine.
I get like 10 shares with a ping of 25 ms and then suddenly there are 1-2 shares with a ping of like 3.000 ms.
I checked my whole shit here in the house - everything works fine.
This must be a problem with ethermine...
Is there anything I can do? Except changing the pool
I get like 10 shares with a ping of 25 ms and then suddenly there are 1-2 shares with a ping of like 3.000 ms.
I checked my whole shit here in the house - everything works fine.
This must be a problem with ethermine...
Is there anything I can do? Except changing the pool
Stale shares can also be caused by the pools stratum being overloaded. Coins like ETH which can have a significant jump in miners can overload the stratum and cause an increased response time, contributing to stale shares. The built in ping utility measures your latency in reaching the server, but it doesn't measure the response time which can cause significant latency if the stratum is overloaded. 2miners have released a tool which also measures the response time from a pools stratum, not just the connection time from your network.
25% stale shares is way too excessive. 5% max is what it should be. I believe it's caused by something in your network, or mining on a shitty pool.
https://2miners.com/blog/check-the-real-ping-to-the-mining-pool-server-with-stratum-ping-tool
https://github.com/2miners/stratum-ping
25% stale shares is way too excessive. 5% max is what it should be. I believe it's caused by something in your network, or mining on a shitty pool.
Quote
Often the performance of various mining pools is verified using the built-in `ping` utility. While this approach is feasible, it is more important to measure the actual response time from the pool, as it may be busy or using some geolocation-based forwarding solution that brings the connecting endpoint closer to the user but then it still takes some significant time for the actual data to be transferred.
Therefore we find useful a mining-specific approach when we measure the amount of time required to connect and successfully pass authentication using the Stratum protocol. This provides more accurate readings in regards of what the mining software will actually do when mining.
We use this tool at 2Miners internally to measure the performance of our pools. It is capable of pinging through IPv4 and IPv6 with or without TLS.
**TL;DR**: The ping to the pool server's box is not as significant as the actual response time through Stratum.
Therefore we find useful a mining-specific approach when we measure the amount of time required to connect and successfully pass authentication using the Stratum protocol. This provides more accurate readings in regards of what the mining software will actually do when mining.
We use this tool at 2Miners internally to measure the performance of our pools. It is capable of pinging through IPv4 and IPv6 with or without TLS.
**TL;DR**: The ping to the pool server's box is not as significant as the actual response time through Stratum.
https://2miners.com/blog/check-the-real-ping-to-the-mining-pool-server-with-stratum-ping-tool
https://github.com/2miners/stratum-ping
Are you using the ethermine pool? recently I also noticed an increase in stale shares, normally it was 1 to 2% but in the last days it is in the range of 7%.
I believe that this is a temporary problem with the pool.
I believe that this is a temporary problem with the pool.
1. I already ordered more cables, heat sinks for the backplates and fans.
2. I don‘t think the wifi is the problem. It works fine for days and then suddenly drops from 1% stale to 30% stale.. that must be a provider issue.
3. There is no possibility to run an ethernet cable to my garage where the miner is located
2. I don‘t think the wifi is the problem. It works fine for days and then suddenly drops from 1% stale to 30% stale.. that must be a provider issue.
3. There is no possibility to run an ethernet cable to my garage where the miner is located
First of all - temperature of course. But yet the problem can be in WLAN - wired network is more stable. If you have opportunity and want to mine seriously - try to change your network.
Thanks Phil, I should have said the GPU temps were the bigger problem.
One other thing, running the 3090 from one cable is at the limit. I hope that PSU has good cables.
If possible, use 2 for the 3090. If you have a less power demanding GPU currently using 2 cables
swap them with the 3090. If you have a second Molex cable you can use it for a riser to free up a
PCIe cable. Just make sure you only use 1 riser per Molex cable. You can also split an 8+6 PCIe
cable to 6+6+6 to power 3 risers with one cable.
When you have to get creative and find alternative solutions to power cabling issues,
use them on the least demanding loads.
Edit: you might also be interested in this thread about Ampere memory temps...
https://bitcointalksearch.org/topic/--5287329
One other thing, running the 3090 from one cable is at the limit. I hope that PSU has good cables.
If possible, use 2 for the 3090. If you have a less power demanding GPU currently using 2 cables
swap them with the 3090. If you have a second Molex cable you can use it for a riser to free up a
PCIe cable. Just make sure you only use 1 riser per Molex cable. You can also split an 8+6 PCIe
cable to 6+6+6 to power 3 risers with one cable.
When you have to get creative and find alternative solutions to power cabling issues,
use them on the least demanding loads.
Edit: you might also be interested in this thread about Ampere memory temps...
https://bitcointalksearch.org/topic/--5287329
LOWER watts a bit and lower ram a bit.
the ram temp numbers you have are very hard on the gear.
try and get some more fans to hit the backs of the gpu's.
Try 900 ram
Try 10 watts less on each gpu.
you will get much better use of the gear if the ram temps drop under 80c
yeah you will lose some hash. but better to lose some hash then lose the gear for an rma.
the ram temp numbers you have are very hard on the gear.
try and get some more fans to hit the backs of the gpu's.
Try 900 ram
Try 10 watts less on each gpu.
you will get much better use of the gear if the ram temps drop under 80c
yeah you will lose some hash. but better to lose some hash then lose the gear for an rma.
Thank you.
It seems like my provider is doing some shitty maintenance....
The ping in my private network is great (even so I use W-Lan)
Luckily it seems like they are done with their shit.
It seems like my provider is doing some shitty maintenance....
The ping in my private network is great (even so I use W-Lan)
Luckily it seems like they are done with their shit.
Stale shares are due to network latency. Some miners report latency, otherwise use ping to
test response time. Don't mine using wifi.
Your temperatures are also too high but that's a different problem.
test response time. Don't mine using wifi.
Your temperatures are also too high but that's a different problem.
Hey guys,
I run a 3090 and a 3080 on Trex for ETH mining.
I run the exact setup for more than 3 days with 2-3 % stale shares.
This evening my stale shares went up to 25%.
I didn't change anything.
Memory temperatures on the 3090 are like 95 celsius all the time. On the 3080 85 celsius all the time.
Any ideas what's going on here and how I can change it?
I already reduced the memory overclock to 950 MHz on both cards. The stale shares went down to like 10 % and then back up to 25% after a short time.
I need to run one riser on molex because I run out of 6/8 Pin cables on the PSU.
And I run the 3090 on one Y shaped 8 Pin from the PSU.
Maybe it's some overheating in the PSU. Could that be a problem? But why after 3 days?
I run a 3090 and a 3080 on Trex for ETH mining.
I run the exact setup for more than 3 days with 2-3 % stale shares.
This evening my stale shares went up to 25%.
I didn't change anything.
Memory temperatures on the 3090 are like 95 celsius all the time. On the 3080 85 celsius all the time.
Any ideas what's going on here and how I can change it?
I already reduced the memory overclock to 950 MHz on both cards. The stale shares went down to like 10 % and then back up to 25% after a short time.
I need to run one riser on molex because I run out of 6/8 Pin cables on the PSU.
And I run the 3090 on one Y shaped 8 Pin from the PSU.
Maybe it's some overheating in the PSU. Could that be a problem? But why after 3 days?
Jump to: