neite99
|
|
August 18, 2014, 11:49:22 PM |
|
Same problem. Anyone know How to handle large size plot with out crashing the miner ? So, I am having a random issue and I am not sure why...
Every once in a while on my server machine, one of the instances that is on my G drive (miner instance) crashes out. Just closes. I don't know why it is doing this. It is only on that one drive, the other miner on the other drive is fine, and the miner on the mapped drive on the same server, mapped to my desktop is fine too. It's only the G drive miner, and randomly...
Any Ideas or suggestions on how to get it to stop? Thanks!
I think my question got overlooked, can anyone help me out with this? Thanks! Is it just the miner that stops or is it a plot generation too? Try adding the word pause on a line after everything else on your bat file java -Xmxblah blah blah blah pauseWhen it dies it will pause the command prompt and ask you to press a button to shut down. This way you can see what the error is that is killing the process. Once you get that it is much easier to help diagnose the problem. Edit: I am assuming your on windows. Yea, I have already added pause to it, and am awaiting the next time it crashes out.... It is only the miner, and only 1 of the 2 miners running on the system. The one on the other drive never crashes. Very strange. The plot generation is fine on both drives. bad sectors of your HDD? I really don't think so... well, I guess, at least I hope not... This drive, as with almost all of the drives/components in the system in question, are almost brand new WD NAS drives. Do you think that bad sectors would cause only the MINER to crash and not the PLOTTER? It would seem to me that it would be the other way around, but I've been known to be wrong before! hehe. How is your NAS attached ? I think you may have misunderstood, or I may have used the wrong context... The drives are NAS drives, but they are physically inside the server, connected like any other drive with SATA and power. Just regular drives, the RED drives from WD, intended for NAS use. This makes it hard to say what is causing the crash. We need the error or check the windows event viewer for anything suspicious. i was/am having this happen to me. started about 3 days ago. running fine for 4.5 days and then random crashes. my single plot size was approach 2TB at the time and my "-Xmx" was limiting it to 750m. I'm wondering whethe it's a mem issue so have upped it to 1024m to see if it makes a diff. i did ask the question here whether there is a log file i can look at the miner logs to to see if it reports why it shutsdown. Im running a drive with 3.5TB (and growing) without the miner dying. Let me get the command for it Edit: Here is my command for the miner: java -Xmx1500m -cp pocminer.jar;lib/*;lib/akka/*;lib/jetty/* pocminer.POCMiner mine " http://127.0.0.1:8125"%* I doubled the -Xmx. It has never crashed. I am using 64 bit java too.
|
|
|
|
Vorksholk
Legendary
Offline
Activity: 1713
Merit: 1029
|
|
August 18, 2014, 11:52:29 PM |
|
For 4TB+ plots, 2GB of RAM for the miner seems to do the trick. Been running steady for 3+ days with no issues.
|
|
|
|
BurstBurst
|
|
August 18, 2014, 11:54:27 PM |
|
2TB and java -Xmx750m . @BurstBurst: What is plot file size and your available RAM and your run_mine command ?
|
|
|
|
neite99
|
|
August 18, 2014, 11:56:29 PM |
|
2TB and java -Xmx750m . @BurstBurst: What is plot file size and your available RAM and your run_mine command ?
try -Xmx1500m
|
|
|
|
BurstBurst
|
|
August 19, 2014, 12:01:41 AM |
|
Trying now. 2TB and java -Xmx750m . @BurstBurst: What is plot file size and your available RAM and your run_mine command ?
try -Xmx1500m
|
|
|
|
burstcoin (OP)
|
|
August 19, 2014, 12:03:15 AM |
|
For 4TB+ plots, 2GB of RAM for the miner seems to do the trick. Been running steady for 3+ days with no issues.
It seems weird for that to be an issue. It shouldn't require more over time unless the gc isn't working right.
|
BURST-QHCJ-9HB5-PTGC-5Q8J9
|
|
|
103deltafox
|
|
August 19, 2014, 12:26:55 AM |
|
BURST for sale... If you want to avoid incurring additional fees on trades and want to complete a basic BTC for BURST exchange PM me... NO HASSLE TRADES
I have 300K + Available... I will even match the AE price of 0.003474 BTC / 1000 BURST
30K lots, 50K lots, 100K lots, and even 300K lots if needed. PM me, Thanx!
Can use escrow, or break it up into smaller tarnsactions and straight trade. Let me know.
That's cheaper than the lowest sell at Poloniex @ 0.00353028 for 52 NXT (which would be required to make that transaction) Dgex is not working for withdrawls or deposits, so actually getting NXT back to BTC has limited options right now...
|
|
|
|
mar28
Newbie
Offline
Activity: 18
Merit: 0
|
|
August 19, 2014, 12:47:42 AM |
|
Wallet can't find recent blocks is it offline ??
|
|
|
|
burstcoin (OP)
|
|
August 19, 2014, 12:51:42 AM |
|
Wallet can't find recent blocks is it offline ?? Nope, just give it a few minutes to connect.
|
BURST-QHCJ-9HB5-PTGC-5Q8J9
|
|
|
tricass
Sr. Member
Offline
Activity: 394
Merit: 250
Crypto enthusiast
|
|
August 19, 2014, 12:57:07 AM |
|
Trying now. 2TB and java -Xmx750m . @BurstBurst: What is plot file size and your available RAM and your run_mine command ?
try -Xmx1500m i had the same settings as u burstburst. i've upped the allowance to 1024m and so far so good. the pc i'm running it on is a win 8.1 with 4bg of ram and running java 1.7.0_60-b19
|
|
|
|
BChydro
|
|
August 19, 2014, 01:07:26 AM |
|
Hey burstcoin, question for you. If I wanted to CPU mine Burst, I understand that I'm limited in speed by the rate I can generate plots, so it's not viable. However, would it be possible to make a mining program that only generated the relevant scoop of each nonce for the current block therefore being 4096x faster?
|
|
|
|
|
burstcoin (OP)
|
|
August 19, 2014, 01:15:43 AM |
|
Hey burstcoin, question for you. If I wanted to CPU mine Burst, I understand that I'm limited in speed by the rate I can generate plots, so it's not viable. However, would it be possible to make a mining program that only generated the relevant scoop of each nonce for the current block therefore being 4096x faster?
No. After the data has reached the full size, the whole buffer is hashed, and the resulting hash is xored with the buffer to finish the plot. So, in order to figure out a scoop you need the pre-xor value of each scoop higher than it, and the hash of all the pre-xor scoops.
|
BURST-QHCJ-9HB5-PTGC-5Q8J9
|
|
|
timk225
|
|
August 19, 2014, 01:27:09 AM |
|
Okay, I am ready to apologize for my earlier assholiness in this topic. But dammit, this coin and miner should have been released with FAR better instructions and details and documentation.
Now that my first miner seems to be working (20 GB partition on a laptop drive while I am on vacation), I am calmer now.
Let's move on. I have serious questions.
|
|
|
|
SpeedDemon13
|
|
August 19, 2014, 01:28:53 AM |
|
Okay, I amready to apologize for my earlier assholiness in this topic. But dammit, this coin and miner should have been released with FAR better instructions and details and documentation.
Now that my first miner seems to be working (20 GB partition on a laptop drive while I am on vacation), I am calmer now.
Let's move on. I have serious questions.
Me and couple other were frustrated too before, so no worries.
|
CRYPTSY exchange: https://www.cryptsy.com/users/register?refid=9017 BURST= BURST-TE3W-CFGH-7343-6VM6R BTC=1CNsqGUR9YJNrhydQZnUPbaDv6h4uaYCHv ETH=0x144bc9fe471d3c71d8e09d58060d78661b1d4f32 SHF=0x13a0a2cb0d55eca975cf2d97015f7d580ce52d85 EXP=0xd71921dca837e415a58ca0d6dd2223cc84e0ea2f SC=6bdf9d12a983fed6723abad91a39be4f95d227f9bdb0490de3b8e5d45357f63d564638b1bd71 CLAMS=xGVTdM9EJpNBCYAjHFVxuZGcqvoL22nP6f SOIL=0x8b5c989bc931c0769a50ecaf9ffe490c67cb5911
|
|
|
SpeedDemon13
|
|
August 19, 2014, 01:30:31 AM |
|
Hey burstcoin, what's the maximum allowance size for a single plot? Or is that only limit to the file system and format?
|
CRYPTSY exchange: https://www.cryptsy.com/users/register?refid=9017 BURST= BURST-TE3W-CFGH-7343-6VM6R BTC=1CNsqGUR9YJNrhydQZnUPbaDv6h4uaYCHv ETH=0x144bc9fe471d3c71d8e09d58060d78661b1d4f32 SHF=0x13a0a2cb0d55eca975cf2d97015f7d580ce52d85 EXP=0xd71921dca837e415a58ca0d6dd2223cc84e0ea2f SC=6bdf9d12a983fed6723abad91a39be4f95d227f9bdb0490de3b8e5d45357f63d564638b1bd71 CLAMS=xGVTdM9EJpNBCYAjHFVxuZGcqvoL22nP6f SOIL=0x8b5c989bc931c0769a50ecaf9ffe490c67cb5911
|
|
|
timk225
|
|
August 19, 2014, 01:31:29 AM |
|
Why are people using 800000 to generate a 200 GB plot? Shouldn't it be 819200 (4096 X 200) to generate a true 200 GB worth of plot nonces? If there's 4096 scoops per bigger thing, wouldn't everything be based off of 4096 instead of 4000?
I saw where someone mentioned Windows 8 having the ability to put multiple hard drives together into one logical drive. Can Windows 7 do this, and is there a name used for the proceedure in Windows?
If I get into this coin, I'd like to purchase a few multi-TB drives and make one huge plot for simplicity. I'm better off not complicating things further than they need to be.
And about pricing this coin for sale - I think since it isn't on any exchange, it should have a set price. X BURST = X BTC, period.
|
|
|
|
drumingspz
|
|
August 19, 2014, 01:35:55 AM |
|
Okay, I am ready to apologize for my earlier assholiness in this topic. But dammit, this coin and miner should have been released with FAR better instructions and details and documentation.
Now that my first miner seems to be working (20 GB partition on a laptop drive while I am on vacation), I am calmer now.
Let's move on. I have serious questions.
Welcome back.
|
░░░░░░░▄▄█████▄▄▄░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░▄██████████████▄░░░░░░░░░░░░░░░░░░░▄██████▄▄░░░░░░ ░░▄██████████████████▄░░░░░░░░░░░░░░▄████████████▄░░░░ ░██████████████████████▄░░░░░░░░░░▄████████████████▄░░ ▄███████████████████████▄░░░░░░░▄████████████████████░ ██████████████████████████▄░░░▄███████████████████████ █████████████████████████████████████████████████████▀ █████████████████████████▀░░░░░▀████████████████████▀░ ░██████████████████████▀░░░░░░░░░▀█████████████████▀░░ ░░███████████████████▀░░░░░░░░░░░░░░▀████████████▀░░░░ ░░░▀███████████████▀░░░░░░░░░░░░░░░░░░▀▀██████▀▀░░░░░░ ░░░░░░▀█████████▀░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
| mira | | ▀▀ ██ ██ ██ ██ ██ ██ ▄▄ | | | ▀▀ ██ ██ ██ ██ ██ ██ ▄▄ | | | |
[/center]
|
|
|
SpeedDemon13
|
|
August 19, 2014, 01:38:03 AM |
|
Why are people using 800000 to generate a 200 GB plot? Shouldn't it be 819200 (4096 X 200) to generate a true 200 GB worth of plot nonces? If there's 4096 scoops per bigger thing, wouldn't everything be based off of 4096 instead of 4000?
I saw were someone mentioned Windows 8 having the ability to put multiple hard drives together into one logical drive. Can Windows 7 do this, and is there a name used for the proceedure in Windows?
And about pricing this coin for sale - I think since it isn't on any exchange, it should have a set price. X BURST = X BTC, period.
Ideally multiples of 4096 would be best because no wasted data or overlapping. Yes you can either do a RAID or a JBOD, a JBOD is the simplest to setup. Also, you can use a GPT to surpass 2TB limitation. I've tried 3 plots files on 3 500GB drives, which resulted in 0 blocks for 4 days of mining. But got 4 blocks when I made a JBOD(Spanning Volume), one 1400GB plot. I got those 4 blocks in the past 24 hours. Maybe just coincidence or maybe the algo runs better on bigger size plots.
|
CRYPTSY exchange: https://www.cryptsy.com/users/register?refid=9017 BURST= BURST-TE3W-CFGH-7343-6VM6R BTC=1CNsqGUR9YJNrhydQZnUPbaDv6h4uaYCHv ETH=0x144bc9fe471d3c71d8e09d58060d78661b1d4f32 SHF=0x13a0a2cb0d55eca975cf2d97015f7d580ce52d85 EXP=0xd71921dca837e415a58ca0d6dd2223cc84e0ea2f SC=6bdf9d12a983fed6723abad91a39be4f95d227f9bdb0490de3b8e5d45357f63d564638b1bd71 CLAMS=xGVTdM9EJpNBCYAjHFVxuZGcqvoL22nP6f SOIL=0x8b5c989bc931c0769a50ecaf9ffe490c67cb5911
|
|
|
BChydro
|
|
August 19, 2014, 01:49:45 AM |
|
Hey burstcoin, question for you. If I wanted to CPU mine Burst, I understand that I'm limited in speed by the rate I can generate plots, so it's not viable. However, would it be possible to make a mining program that only generated the relevant scoop of each nonce for the current block therefore being 4096x faster?
No. After the data has reached the full size, the whole buffer is hashed, and the resulting hash is xored with the buffer to finish the plot. So, in order to figure out a scoop you need the pre-xor value of each scoop higher than it, and the hash of all the pre-xor scoops. Very clever
|
|
|
|
|