How often do your Rigs restart or crash?

smoolae

full member

Activity: 602

Merit: 106

Quote from: Spllouder on January 31, 2018, 12:23:25 AM

Well!!

Seems adding the 600B did the trick! well its been :

ETH - Total Speed: 88.754 Mh/s, Total Shares: 831, Rejected: 0, Time: 27:02

27 hours since i added the 2nd PSU, I will get you updated if it happen's again.

BTW what i discovered on inspecting, GPU1 this was the only SATA wires warm to touch which gave the error, Also was the second gpu on sata from PSU, So now only one GPU per Sata line to relive any stress on the wires.

Yeah, be careful with using SATAs. You can overload them really easily without you noticing until crash, smoke and flames Tongue

.

Spllouder

newbie

Activity: 24

Merit: 0

Well!!

Seems adding the 600B did the trick! well its been :

ETH - Total Speed: 88.754 Mh/s, Total Shares: 831, Rejected: 0, Time: 27:02

27 hours since i added the 2nd PSU, I will get you updated if it happen's again.

BTW what i discovered on inspecting, GPU1 this was the only SATA wires warm to touch which gave the error, Also was the second gpu on sata from PSU, So now only one GPU per Sata line to relive any stress on the wires.

tinfoilMining

newbie

Activity: 28

Merit: 0

Quote from: tinfoilMining on January 29, 2018, 10:39:15 PM

Restarts once every hour on average on my only dedicated rig Grin

It's been pulling 650-660 watts from wall on a 750 gold PSU, so maybe that could be the culprit. I reduced power target after reading the posts above. Now pulling 610 watts, we'll see how that works out.

13,5 hours and still no crash yet.. Wow could it be that simple. I was certain it was the risers, and was going to change and test every riser to see if that was the problem.

This thread can have saved me allot of headache Cheesy

x8664amd

jr. member

Activity: 129

Merit: 6

Long term HODLer since 2014

Never crashes, but I tend to just restart them in batches at the beginning of each month.

lathrodectus

newbie

Activity: 25

Merit: 0

I get no crashes or reboots for months. Unless i reboot them Cheesy

kaguraqt

newbie

Activity: 2

Merit: 0

No, my rig never crash. maybe you rig restarted because your Mod bios or overclocking too much

manji

full member

Activity: 210

Merit: 174

BookiePro.Fun - The World's Betting Exchange

Quote from: Spllouder on January 29, 2018, 11:49:45 AM

Well, I think i finally got my system stable. I swapped out MB and now using H270A Pro. Before i had allot of Crash restarts, until now it is very stable, besides last night. The rig never turned off, Fans where same speed as always. But the room was cold, I turned on the mointer and could not get any signal. Checked Nanopool, and it was not hashing for 3 hours. Any idea's ?

So my question to everyone is how often your Rig crashes and what was the cause and fix?

My rig never rested 24/7 had to work for me, rarely Crash because, before mining, I check hardware condition, air flow condition for the room mining, calculate the power consumptions for CPU and GPU, check compatibility mining software. if everything good, then mining do it and I'm monitor for 3 or 4 hours.

if have crashed, when power electricity down, switch to new coins, that's all. atm, it's been 3 months Rig never crash.

blottominer

newbie

Activity: 1

Merit: 0

Quote from: KaydenC on January 29, 2018, 02:04:43 PM

Mine depends on the algo. For zcoin I have about 7 crashes a day out of 15 rigs. For equihash it drops to 2-3 crash a day out of 23 rigs. They all have scripts that detects idle gpus and restarts the PC so it's automated. Likely my zcoin core oc is too high though.

Hi there, just curious if you wrote the scripts yourself or if there are readily available ones out there.

I am now up to 6rigs with 8x1070ti each, and they seem to run into random issues around the 18-36 hour range. A restart fixes it, would love to get that part of the operation automated.

Thanks!

KaydenC

sr. member

Activity: 610

Merit: 265

Quote from: VyprBTC on January 29, 2018, 09:05:41 PM

Quote from: KaydenC on January 29, 2018, 02:04:43 PM

Mine depends on the algo. For zcoin I have about 7 crashes a day out of 15 rigs. For equihash it drops to 2-3 crash a day out of 23 rigs. They all have scripts that detects idle gpus and restarts the PC so it's automated. Likely my zcoin core oc is too high though.

Trying to dig up your scripts, used to be in your sig but can't find it - Phi is giving me a lot of memory error crashes so that would have been cool to use, changed algos though couldn't deal with it.

Oh I didn't post it before.

Basically I use nvidia SMI to poll gpu power draw every 10 mins and write that to a text file. It then loops through that text file, if any value below 50W, restart the PC.

With this automation, it's possible to play more aggressive with the clocks often yielding 1-3% improved hash rates at the same power draw despite more crashes.

tinfoilMining

newbie

Activity: 28

Merit: 0

Restarts once every hour on average on my only dedicated rig Grin

It's been pulling 650-660 watts from wall on a 750 gold PSU, so maybe that could be the culprit. I reduced power target after reading the posts above. Now pulling 610 watts, we'll see how that works out. I thought I was way within stable territory at 650..

My gaming/mining rig is very stable though, but that's only two GPU's and no risers ..

KK120

newbie

Activity: 33

Merit: 0

I have 6 AMD 470 8GB Rig running windows 10 and STAK-XMR-AMD, since one month+ now.

It got stack a couple of times in the beginning, every 4-5 days but so far it worked non stop for the last 2 weeks until another freeze as we speak Sad

VyprBTC

full member

Activity: 350

Merit: 100

Quote from: KaydenC on January 29, 2018, 02:04:43 PM

Mine depends on the algo. For zcoin I have about 7 crashes a day out of 15 rigs. For equihash it drops to 2-3 crash a day out of 23 rigs. They all have scripts that detects idle gpus and restarts the PC so it's automated. Likely my zcoin core oc is too high though.

Trying to dig up your scripts, used to be in your sig but can't find it - Phi is giving me a lot of memory error crashes so that would have been cool to use, changed algos though couldn't deal with it.

CarlOrff

sr. member

Activity: 826

Merit: 440

My RIG never crash, but I se Ubuntu and nVidia GPUs.
I restart it + or - once a month for cleaning the dust.
In the past, I had 2 serious issues but it was with the mining software (when I used a new one).

Spllouder

newbie

Activity: 24

Merit: 0

Quote from: 2stout on January 29, 2018, 01:12:05 PM

What is the total wattage draw of the GPU's & MB and the rating of the PSU(s)? As a rule of thumb, you don't want to exceed 80% of the PSU capacity. Also may need to check circuit this on and make sure it isn't being pushed beyond 80% -ex, if on 15A ckt, then W= A x V, assuming 120V, then Watts=1800 and 80% of this= 1440W.

Great stuff here!

I forgot to mention i have a EVGA 750 G2 Gold

I pulling 632.5 @ this moment typing from the wall.

I have a EVGA 600B Bronze ready to add, but waiting for pcie extension cable's.

When i do get 6 gpu's, I will run four RX570 just on 750G2 and the other 2 RX570's and System on the 600B.

ALSO?

I use at the moment Sata to 6pin to power my risers, i have molex to 6pin in mail. It has two molex into 6pin, do i need to hook up both molex's to work?

Ok i just did the math, 80 % of 750 = 600

I'am pulling 615 to 630. I need to add this 600B ASAP!

logosobscura

member

Activity: 84

Merit: 16

Tru Reputation Network Pre-Sale: tru.ltd/tokensale

As everyone has said- check the PSUs- 5 years of experience has taught me not to cheap out on PSUs and that the math isn't linear. Cheap PSUs are essentially a waste of money, and more than GPUs they're where you should put the majority of your planning. Also applies to PCI-E risers/extensions- don't cheap out on them if you're going to stack on a motherboard and A/B test as you scale up.

smoolae

full member

Activity: 602

Merit: 106

Depends on an algorithm, Cryptonight runs stable af. Ethash and Equihash do give me some problems.

martyroz

full member

Activity: 325

Merit: 110

Active workers

Code:

XNSUB  	WORKER  ALGORITHM  SPEED (ACCEPTED)  SPEED (REJECTED)  DIFF  TIME (MIN)  LOCATION  
Yes	041	Equihash	4177.86 Sol/s	0 Sol /s (0%)	4096	2267	USA
Yes	143	Equihash	3597.6 Sol/s	0 Sol /s (0%)	4096	655	USA
Yes	241	Equihash	5106.27 Sol/s	0 Sol /s (0%)	4096	2268	USA
Yes	VEGA56	CryptoNight	10.67 kH/s	0 kH /s (0%)	400000	2959	HK

looks like 40+ hours for three rigs and 11 hours for 1 rig
I have one rig down ATM - can't check till I get home. It's running Windows + Vega and is a PITA...

MarkAz

hero member

Activity: 687

Merit: 511

We have several hundred rigs - it's rare for them to crash or reboot, generally speaking that's an indicator of some secondary issue. Typically we'll have one system per day or two that needs a reboot for whatever reason, but as a percentage that's like 0.1 % which is more than acceptable to us. The big thing is making sure you have some way to remotely manage things, that will make you life alot easier if you start scaling up.

Oh, and we're running Linux for everything, so you might get different results depending on what OS you choose. I have my gaming box mine when I'm not playing games, and normally it only rebooted when a Windows update happened - but now after that stupid Intel bug fix it probably reboots about once every day or two.

KaydenC

sr. member

Activity: 610

Merit: 265

Mine depends on the algo. For zcoin I have about 7 crashes a day out of 15 rigs. For equihash it drops to 2-3 crash a day out of 23 rigs. They all have scripts that detects idle gpus and restarts the PC so it's automated. Likely my zcoin core oc is too high though.

2stout

hero member

Activity: 2464

Merit: 603

What is the total wattage draw of the GPU's & MB and the rating of the PSU(s)? As a rule of thumb, you don't want to exceed 80% of the PSU capacity. Also may need to check circuit this on and make sure it isn't being pushed beyond 80% -ex, if on 15A ckt, then W= A x V, assuming 120V, then Watts=1800 and 80% of this= 1440W.

Topic: How often do your Rigs restart or crash? (Read 360 times)