A4 Dominator - Pre-Order Group Buy - 280mh, roughly 1000w, $1800 + shipping

[vinylwasp]

Network would be different as well. Might be useful to do some analyzing of connection latency as well. I'm not sure how the A4 work but perhaps if it thinks there is network issues it reboots as well?

Yes, check the 'Get Failures' stats for Getwork failures first, and then look at Pool Rejects for high Pool stales.

[Eyedol-X]

Network would be different as well. Might be useful to do some analyzing of connection latency as well. I'm not sure how the A4 work but perhaps if it thinks there is network issues it reboots as well?

Yes, check the 'Get Failures' stats for Getwork failures first, and then look at Pool Rejects for high Pool stales.

Think I figured it out.

Had to SSH into the PI and clear the tmp directory which contained a cgminer.lock file

Once I cleared that, it appears to be running normally again... almost 10 30 mins now without a reboot and prior it would reboot around 5 minutes like clock work.

Owners of the A4, I recommend you take note of this because if it happened to me, its bound to happen to you -- All I did was power it off and move it.

Questions:
1. what is the function of cgminer.lock?
2. what event possible triggers occurrence this cgminer.lock?
3. do we need to clear tmp directory regularly?

Best I can tell it's a temp file that gets created that contains information while the miner is running.

From my research, when the raspberry pi reboots, its supposed to clear the tmp directory. However when you just flip the power switch and shut it off, the Pi doesn't have the opportunity to clear this file. I only recommend doing this if you're having issues getting your A4 to mine or if its in a 5 minute reboot loop like mine was.

I took a guess and manually cleared the file/directory by performing the following steps.

1. Make SURE the miner is stopped (note: I had issues with the thing starting itself up even with Manual Start configured)
2. SSH Into the Pi
3. Navigate to tmp directory in the root
4. Clear contents of tmp directory with command

Code:

sudo rm -r *

5. Restart the miner

I observed this file was re-created when the miner started again.

Over 16 hours now since I did this and its working fine again.


If you notice I also set the time/date correctly on the Pi finally so that's accurate in this screen capture.

[Eyedol-X]

Update:

I have shipped the Jig on to the next person.

When I received the Jig, the card inside had become detached and was only being held by 1 corner of glue. At the direction of Innosilicon I re-secured the Jig before shipping it to the next person.

I used specific flexible adhesive hot glue and not regular hot glue as the hot glue previously used got hard and had come off during the 1st shipment.

Images of the repair are in the Jig Reflash Thread: https://bitcointalk.org/index.php?topic=1718024.msg17472203#msg17472203

[Eyedol-X]

Network would be different as well. Might be useful to do some analyzing of connection latency as well. I'm not sure how the A4 work but perhaps if it thinks there is network issues it reboots as well?

Yes, check the 'Get Failures' stats for Getwork failures first, and then look at Pool Rejects for high Pool stales.

Think I figured it out.

Had to SSH into the PI and clear the tmp directory which contained a cgminer.lock file

Once I cleared that, it appears to be running normally again... almost 10 30 mins now without a reboot and prior it would reboot around 5 minutes like clock work.

Owners of the A4, I recommend you take note of this because if it happened to me, its bound to happen to you -- All I did was power it off and move it.

Questions:
1. what is the function of cgminer.lock?
2. what event possible triggers occurrence this cgminer.lock?
3. do we need to clear tmp directory regularly?

Best I can tell it's a temp file that gets created that contains information while the miner is running.

From my research, when the raspberry pi reboots, its supposed to clear the tmp directory. However when you just flip the power switch and shut it off, the Pi doesn't have the opportunity to clear this file. I only recommend doing this if you're having issues getting your A4 to mine or if its in a 5 minute reboot loop like mine was.

I took a guess and manually cleared the file/directory by performing the following steps.

1. Make SURE the miner is stopped (note: I had issues with the thing starting itself up even with Manual Start configured)
2. SSH Into the Pi
3. Navigate to tmp directory in the root
4. Clear contents of tmp directory with command

Code:

sudo rm -r *

5. Restart the miner

I observed this file was re-created when the miner started again.

Over 16 hours now since I did this and its working fine again.


If you notice I also set the time/date correctly on the Pi finally so that's accurate in this screen capture.

Update on this, ran fine for 3 days then yesterday around 5am this problem started again. I didn't catch it until today. I again deleted the cgminer.lock file and rebooted the pi manually as mentioned above and it seems to have corrected the problem for now...

Lost 24 hours of mining because I wasn't watching it...

I guess next step is to take another copy of 2.0.2 on an SD card and go swap it out.

Anyone know of a good way of changing the default IP address without a 2nd Pi to change it from? I can see the necessary files in the image but can't find a successful method of editing it from Windows... maybe it can be done from Linux?

[Chazaki]

My a4 will hash great at rated speed for a couple days and then speed drops by almost 100 mhs. Not sure the cause...

[kurbeks]

Pool problems, Dropped boards? Nicehash?

[Hhampuz]

So what's the average for users now? In terms of speed, uptime and stability? Does anyone think inno might lower the prices considering bitmains new miners seem very solid and stable? I'm leaning more towards bitmain but always open to get persuaded back to inno

[kurbeks]

If L3 is anything like S9 you gonna all what you saved for reapairs and shipping miner back to China. When thinking about Scrypt vs SHA256 and ETH/equi Scrypt is least affected by electricity prices, so unless you mine just LTC you can gain money even with elect prices. So L3 gain over A4 is much smaller than for example S9 vs it's competitors.

If L3 would like 1kW and 450MHash/s that would have sold me, now i don't think so.  A4 still pumping hard.

[Hhampuz]

If L3 is anything like S9 you gonna all what you saved for reapairs and shipping miner back to China. When thinking about Scrypt vs SHA256 and ETH/equi Scrypt is least affected by electricity prices, so unless you mine just LTC you can gain money even with elect prices. So L3 gain over A4 is much smaller than for example S9 vs it's competitors.

If L3 would like 1kW and 450MHash/s that would have sold me, now i don't think so.  A4 still pumping hard.

I get that, but there's still be problem with stability, there has been a lot of problems throughout this thread with cards dropping, getting up to 50Mh/s less than advertised, close to no response from inno on these issues etc. I know that there is the problem with help/spare parts with bitmain, but so far this does not seem like that much of a better choice in those departments.

[kurbeks]

After reflash or B3 everything works more or less ok. Worse case need hard restart takes 30 secs to start up again. At least no boards have died yet.

[Searing]

Supposedly the A4's dropped in price...or so the site says. They don't list it. Anyone know yet. *just curious don't plan
on getting any*

[kurbeks]

Probably to 1780 or something

[Searing]

Probably to 1780 or something

Heh..they are soooooo clueless. (Hell, what do I know they may be selling them to newbies like hotcakes)

You pretty much nailed it. Here is their reply

Hello,

The clearance sale of A4 before Jan 21st is $1700（$100 off the list price） per unit plus shipping fee.

A WHOPPING 100 bucks OFF the price of ..what was it 3 months ago?

sheesh..it is either they are selling very few...or selling lots to a newbie crowd..I see no real in between.

[Inno_ASIC]

A4, 280Mhs, the most powerful miner in the market! Last batch clearance price before 21st Jan.! Buy from http://www.innosilicon.com/html/a4-miner/

[Hhampuz]

Probably to 1780 or something

Heh..they are soooooo clueless. (Hell, what do I know they may be selling them to newbies like hotcakes)

You pretty much nailed it. Here is their reply

Hello,

The clearance sale of A4 before Jan 21st is $1700（$100 off the list price） per unit plus shipping fee.

A WHOPPING 100 bucks OFF the price of ..what was it 3 months ago?

sheesh..it is either they are selling very few...or selling lots to a newbie crowd..I see no real in between.

I have to agree with you. My guess would be that they are not selling nearly enough and with the $100 off (lol) I'd guess we won't see batch 2 too soon. Pretty sad really, because ad you say it seems there are newbies buying these units..

[Eyedol-X]

Network would be different as well. Might be useful to do some analyzing of connection latency as well. I'm not sure how the A4 work but perhaps if it thinks there is network issues it reboots as well?

Yes, check the 'Get Failures' stats for Getwork failures first, and then look at Pool Rejects for high Pool stales.

Think I figured it out.

Had to SSH into the PI and clear the tmp directory which contained a cgminer.lock file

Once I cleared that, it appears to be running normally again... almost 10 30 mins now without a reboot and prior it would reboot around 5 minutes like clock work.

Owners of the A4, I recommend you take note of this because if it happened to me, its bound to happen to you -- All I did was power it off and move it.

Questions:
1. what is the function of cgminer.lock?
2. what event possible triggers occurrence this cgminer.lock?
3. do we need to clear tmp directory regularly?

Best I can tell it's a temp file that gets created that contains information while the miner is running.

From my research, when the raspberry pi reboots, its supposed to clear the tmp directory. However when you just flip the power switch and shut it off, the Pi doesn't have the opportunity to clear this file. I only recommend doing this if you're having issues getting your A4 to mine or if its in a 5 minute reboot loop like mine was.

I took a guess and manually cleared the file/directory by performing the following steps.

1. Make SURE the miner is stopped (note: I had issues with the thing starting itself up even with Manual Start configured)
2. SSH Into the Pi
3. Navigate to tmp directory in the root
4. Clear contents of tmp directory with command

Code:

sudo rm -r *

5. Restart the miner

I observed this file was re-created when the miner started again.

Over 16 hours now since I did this and its working fine again.


If you notice I also set the time/date correctly on the Pi finally so that's accurate in this screen capture.

Update on this, ran fine for 3 days then yesterday around 5am this problem started again. I didn't catch it until today. I again deleted the cgminer.lock file and rebooted the pi manually as mentioned above and it seems to have corrected the problem for now...

Lost 24 hours of mining because I wasn't watching it...

I guess next step is to take another copy of 2.0.2 on an SD card and go swap it out.

Anyone know of a good way of changing the default IP address without a 2nd Pi to change it from? I can see the necessary files in the image but can't find a successful method of editing it from Windows... maybe it can be done from Linux?

Another Update.

So after about 2-3 Hours of working normally after this post, the A4 went down again and started boot looping every 5 mins or so. I didn't have opportunity to check on it overnight and the next morning I found it completely offline and the Pi unable to be reached via ping, ssh, or accessible via web interface.

I power-cycled the entire unit and it came back online normally. Within 5 minutes, it rebooted again. I got the A4 back up and stopped mining. I changed the speed to 1000mhz and started it mining, rebooted again within 5 mins.

I want to stress this is the Pi rebooting for what appears to be a script going off, the A4 appears to be hashing normally when it happens.

I reapplied all the settings, set everything to 1200mhz, typed back in my pool information, added a 3rd pool and started mining again.

Now its be running for over 9 hours on PH without an issue. 

I've got new SD cards coming today and I'm going to replace the one in the Pi with a fresh 2.0.2 image/sd card the next time it goes down again.

I'm fairly certain at this point that I have not seen the end of this issue.

I reached out to Inno for assistance in steps I can take if this issue persists after I replace the SD Card / Image.

[kurbeks]

Looks like Pi going nuts. Did you try reseting it with  button? If new SD card and PI reset doesnt help, probably PI broken.

[Inno_ASIC]

Network would be different as well. Might be useful to do some analyzing of connection latency as well. I'm not sure how the A4 work but perhaps if it thinks there is network issues it reboots as well?

Yes, check the 'Get Failures' stats for Getwork failures first, and then look at Pool Rejects for high Pool stales.

Think I figured it out.

Had to SSH into the PI and clear the tmp directory which contained a cgminer.lock file

Once I cleared that, it appears to be running normally again... almost 10 30 mins now without a reboot and prior it would reboot around 5 minutes like clock work.

Owners of the A4, I recommend you take note of this because if it happened to me, its bound to happen to you -- All I did was power it off and move it.

Questions:
1. what is the function of cgminer.lock?
2. what event possible triggers occurrence this cgminer.lock?
3. do we need to clear tmp directory regularly?

Best I can tell it's a temp file that gets created that contains information while the miner is running.

From my research, when the raspberry pi reboots, its supposed to clear the tmp directory. However when you just flip the power switch and shut it off, the Pi doesn't have the opportunity to clear this file. I only recommend doing this if you're having issues getting your A4 to mine or if its in a 5 minute reboot loop like mine was.

I took a guess and manually cleared the file/directory by performing the following steps.

1. Make SURE the miner is stopped (note: I had issues with the thing starting itself up even with Manual Start configured)
2. SSH Into the Pi
3. Navigate to tmp directory in the root
4. Clear contents of tmp directory with command

Code:

sudo rm -r *

5. Restart the miner

I observed this file was re-created when the miner started again.

Over 16 hours now since I did this and its working fine again.


If you notice I also set the time/date correctly on the Pi finally so that's accurate in this screen capture.

Update on this, ran fine for 3 days then yesterday around 5am this problem started again. I didn't catch it until today. I again deleted the cgminer.lock file and rebooted the pi manually as mentioned above and it seems to have corrected the problem for now...

Lost 24 hours of mining because I wasn't watching it...

I guess next step is to take another copy of 2.0.2 on an SD card and go swap it out.

Anyone know of a good way of changing the default IP address without a 2nd Pi to change it from? I can see the necessary files in the image but can't find a successful method of editing it from Windows... maybe it can be done from Linux?

Another Update.

So after about 2-3 Hours of working normally after this post, the A4 went down again and started boot looping every 5 mins or so. I didn't have opportunity to check on it overnight and the next morning I found it completely offline and the Pi unable to be reached via ping, ssh, or accessible via web interface.

I power-cycled the entire unit and it came back online normally. Within 5 minutes, it rebooted again. I got the A4 back up and stopped mining. I changed the speed to 1000mhz and started it mining, rebooted again within 5 mins.

I want to stress this is the Pi rebooting for what appears to be a script going off, the A4 appears to be hashing normally when it happens.

I reapplied all the settings, set everything to 1200mhz, typed back in my pool information, added a 3rd pool and started mining again.

Now its be running for over 9 hours on PH without an issue. 

I've got new SD cards coming today and I'm going to replace the one in the Pi with a fresh 2.0.2 image/sd card the next time it goes down again.

I'm fairly certain at this point that I have not seen the end of this issue.

I reached out to Inno for assistance in steps I can take if this issue persists after I replace the SD Card / Image.

Hi Eyedol,

From the feedback of our customer who use v2.0.2beta, the performance of the A4 is stable without reboots issues. So something is wrong here.

Just curious that you can access to the SSH because till now we didn't open source. Please DON'T CHANGE CODE RANDOMLY. Because we don't know what will happen when you changed the code on your side.

For now, please get the Pi restarted first then try the new SD card.
We are glad to help if you need.

Inno Support Team

[Inno_ASIC]

Update:

I have shipped the Jig on to the next person.

When I received the Jig, the card inside had become detached and was only being held by 1 corner of glue. At the direction of Innosilicon I re-secured the Jig before shipping it to the next person.

I used specific flexible adhesive hot glue and not regular hot glue as the hot glue previously used got hard and had come off during the 1st shipment.

Images of the repair are in the Jig Reflash Thread: https://bitcointalk.org/index.php?topic=1718024.msg17472203#msg17472203

Thanks very much, Eyedol. The next person has received the Jig yesterday.
Really appreciate your maintenance for the Jig 

Update of the Jig delivery in USA:
1. the next delivery of Jig will be finished soon for the rest 4 users
2. we will ship another Jig from Europe to USA soon

[firehawk71]

My graph After 7 days mining non stop 3 cubes total on litecoinpool