If your app is periodically establishing connection, but then instead of actually mining on them, just terminating them, then you get yourself periodically banned for doing this.
We have strict anti-DDOS rules and such behaviour qualifies as an attack.
well this is unfortunate, how to check the stratum availability without triggering those bans? unfortunately most miner developers dont want to implement failover (except sgminer, but only conf files afaik) |
|
|
|
wenn sie jemand alle auf einmal nimmt lässt sich bestimmt noch was machen  |
|
|
|
Well anything electrically powered can catch up the fire. Sometimes, you do have to risk some stuff.
If you ask me, I'd never worry about the fire and let the system be powered. All the new GPU's are very much safe in this regard, in the worst case, your GPU's might burn, your cables might burn, your motherboard might burn, but your home, nah, I don't think its possible. There may be a single case in a million rigs like this, but I don't think that anyone should worry about this, specially if you have latest hardware.
actually its quite possible if even a small phone can burn down a house probably safer to not use wooden frames and have nothing stacked on top |
|
|
|
im pretty sure it had to do with dust buildup, ... Dust is never conductive to those low DC voltages. Dust would never create that spark. However a stray metal fragment might. Normal is for electronics to be heavily caked in dust without failure or arcing. More likely something internally failed inside a semiconductor creating a physic defect. Most all electronic failures have no indication. But some rare ones (such as this one) do. So a short circuit caused the PSU to enter current foldback limiting - and shutdown. Later, the same failure was now a tiny short circuit that burned open during a new power up only creating smoke. Not flame as originally suggested. Just arcing and smoke. A major difference. thanks for the explanation! there was smoke and iirc there also was a small flame at first immediately after power on but well, i buried the card long ago, may it rest in piece |
|
|
|
i had one card create a spark, the psu cut the power immediately i was almost asleep at that time and sleepy as i was i thought "wut?" and turned it back on, at this point the gpu really burned and created smoke and the psu would not power down automatically in my case nothing could have prevented that from happening besides me cleaning the card, it was dust shorting something inside the card  You remind me of some cards i had that i didn't clean for decades  until the dust became a part of it and those cards had the longest life i have ever seen for gpus 6 years at minimum...i think it's not the dust or anyother fault, electrical components in the gpus can go dead in silence or it can also go dead after it burns some fire and smoke.....shit happens , by the way dust can't make a short circuit because of it's high resistance. im pretty sure it had to do with dust buildup, the card in question was a 7970 with an arctic cooling cooler with this special backplate. betweet backplate and actual back of the card there is a very small gap, and dust was all over it on the edges (where the fire and smoke came out), might have been something else as well, but well, first and only card to die on me for now |
|
|
|
with my stratum checker now available on my manager i was able to observe that many stratum servers are not responding from time to time for longer than 10 sec when initially connecting, which results in a timeout on my checker:  i have compared that to the output of some claymore eth miner running at that time where it reported "timeout", it continued to work flawlessly. i have verified not a single response is sent on the first 10 sec, so my guess here is that existing connections work fine, but new ones are getting blocked/rejected/dropped or something from time to time there really is a problem so the checker works as expected as the miner itself also fails to connect, but these cases seem rare (besides lbry eu stratum which is kinda buggy) maybe you guys know why this is happening? |
|
|
|
I've discovered cryptonight is broken in 3.5.5, use 3.5.4 instead. I'm looking into it.
Got unlucky, first submit rejected and panicked. I made no changes in cryptonight so I was scared it
was another strange bug.
Cryptonight looks good.
Haha, good to hear I'll send some small donation later, get youself a drink |
|
|
|
i had one card create a spark, the psu cut the power immediately i was almost asleep at that time and sleepy as i was i thought "wut?" and turned it back on, at this point the gpu really burned and created smoke and the psu would not power down automatically in my case nothing could have prevented that from happening besides me cleaning the card, it was dust shorting something inside the card |
|
|
|
also my amds are now running past 50 accepted shares on timetravel on windows, so not submitting the benchs prevented the issue
Have you tried submitting? -p stats just tried, still exits on stats enabled |
|
|
|
|
also my amds are now running past 50 accepted shares on timetravel on windows, so not submitting the benchs prevented the issue
|
|
|
|
|
thanks for this
small note: the windows bin for sse2 was renamed to *-core2.exe instead of *-sse2.exe as in previous builds, i renamed it to sse2 in my miner repo to not break scripts already in place for previous versions
cheers
|
|
|
|
@ioglnx
we got offtopic
you know CPU usage on opencl pascal is due no one tested it on NV and fixed the bug? also pascal mining was not profitable on NV at all, and the wallet/coin is so big FUCKUP it prolly will never be again
Offtopic you complained about your weak gpu..i just told you that with the proper cpu you wont have this issue. And this had nothing to do with the opencl at all..it is one piece of the puzzle. but looks like you wanna have the last freakin word of hot air. Ok you are right, and still I am mining zcash, decret, cuda pascal, lbry, ethereum, FTC, + CPU mining at same time on rig of 5 x 1070 on dual core CPU with no issue at all Because GPU miner should not and does not use CPU more then few % pascal opencl for NV is broken, no one fixed it or try it to fix CPU usage, its stupid to have 8 core CPU on 6 GPU rig due same stupid bug obviously you all should have gotten amd cards, then there would be no issue with the cpu load /sarcasm |
|
|
|
|
i hope all people mining on zpool are aware of the 20% income stealing?
read the last 10 pages or so of the original zpool ann, which is now closed, if in doubt
|
|
|
|
@ ioglnx I can only say LOL! I would invest 8 core CPU for my rigs if any benefit can be faund (expect the bugs)  But you see where u end up if you by cheap and not properly doing your math. Its well known about resources required by mulit gpu setup. In the end ..if I were you I would cry..or stop cry and accept the fact i saved on the wrong part. *cough* great read, haha a celeron/pentium is totally fine for multi gpu rigs, though |
|
|
|
working great, now only waiting for nhm integration as im using that for my nvidia cards |
|
|
|
They have also a miner called Excavator 1.0.0a i don't know if it is nvidia only but they showed it in the pascal thread
interesting, seems nvidia only and probably based on eqm/nheqminer from the looks my initial question was regarding amd gpus, i hope there is a fix soon edit: nevermind, its working well on amd multi card rigs if supplied with the magic extra params also noted on suprnova |
|
|
|
PascalNEW algorithm (for PascalCoin) added to NiceHash marketplace. Sellers connect your miners to: - stratum+tcp://pascal.eu.nicehash.com:3358
- stratum+tcp://pascal.usa.nicehash.com:3358
- stratum+tcp://pascal.hk.nicehash.com:3358
- stratum+tcp://pascal.jp.nicehash.com:3358
NiceHash Miner users will also be able to auto-switch to Pascal when we update our software. Buyers can order hashing power here: https://www.nicehash.com/?p=orders&a=25 i would love to use it but yours and ocminers sgminer produce a very high cpu load thus making mining impossible on multicard rigs, any eta? |
|
|
|
Hi, this software has been around for some time, but i have not announced it yet, so here it is Miner-Manager: - add groups, miners and pools
- monitor miners via api stats if available
- autoswitching pools (only nicehash currently)
- stratum availability checks
- pool priorities per group with failover if stratum not working
- written in nodejs, update and reboot scripts for linux and windows (linux *should* work, however im using it only for cpu mining)
Github: https://github.com/felixbrucker/miner-managerthis is beta software, it might error, i just did some rework and there are probably some bugs to hunt from that it now uses a stratum checker to enable failover with priorities for generic mining software as well as generic pool config for known miners (others will still have to type in their full cmdline) as of now there is only a workaround if your miner does not connect to a stratum pool (you will need to create a dummy pool which is online and select that one as pool for the group) but is a standalone miner or something, will see when/how to ease that also i have reduced the algos, if you require any just tell, i have added only the major ones Disclaimer: im coding it mainly for myself/my needs, but however everyone is free to use it and modify it cheers |
|
|
|
|
nheqminer doesnt support multiple pools and failover strategies (yet)
however im currently writing a small external program which will detect if a stratum server is offline/not working
im planning to use this in my manager to implement backup pools, but this will eventually take a while, im currently studying for my exams (at least i should)
|
|
|
|
|
Hi,
im currently writing a small program to check stratum availability and authorization status and it works for some algos i have tested at nicehash, even ssl/tls works fine
only cryptonight doesnt seem to work
i have read that the cryptonight "stratum" is stratum like and i have seen that there is a "login" param instead of "method" where you just seem to skip the subscribe part and go straight to "authorize"
im getting the following responses:
nicehash plain socket: connected and immediately closed after sending the req
nicehash ssl socket: connected and immediately closed after sending the req
miningpoolhub plain socket: connected and mostly timeout after sending req, sometimes it works and i get a job
miningpoolhub ssl socket: connected and mostly timeout after sending req, sometimes it works and i get a job
req im sending: req = '{"id":2, "jsonrpc":"2.0", "login":"'+worker+'", "pass": "'+pass+'", "agent": "stratumTest"}';
(note: \n gets appended afterwards)
what am i doing wrong here?
edit: ok nevermind, found the solution in the cpuminer-opt sourcecode:
req = '{"id":2, "jsonrpc":"2.0", "method":"login", "params": {"login":"'+worker+'", "pass": "'+pass+'", "agent": "stratumTest"}}';
|
|
|
|
|
Show Posts
