HASHRA CONTROLA on Raspberry Pi for Gridseed - page 7.

vaudvaud

member

Activity: 60

Merit: 10

Quote from: dloganbill on April 24, 2014, 11:02:39 PM

I seem to have a lot of HW issues with the latest firmware. I've been mining at Nicehash @850. I have two sets of 10 miners running on the hardware Hashra sent. 10 miners per rpi. Any ideas on what is causing this?

I had this happen twice to me... once it was a bad connection to one of the Gridseeds, after resetting all the cables it fixed it... the other time it ended up being a gridseed that went bad, took me awhile to find out which one was throwing all the hardware errors (which is why I wish they showed HW errors per each device in addition to the submitted/accepted).. once I replaced it I've been fine. I'm running 12 off a single powered usb hub @ 850. I also have mine restart the miner every 6 hours. Been rock solid stable the last few days... which is good since I'm going to be gone the next 3 days and I hope they keep going no problems during that time.

fivejonnyfive

sr. member

Activity: 294

Merit: 250

Quote from: bexwhitt on April 25, 2014, 08:45:57 AM

Quote from: maxzilla on April 25, 2014, 04:25:00 AM

Quote from: bexwhitt on April 25, 2014, 03:21:34 AM

1.4 runs for a few hours then you can't connect to the web interface, I get

Forbidden

You don't have permission to access / on this server.

if you look at the raspberry pi output it's doing this

https://www.youtube.com/watch?v=48bALCHC0qI&feature=youtu.be

the miners seem to be still working though

hmm, it looks like the controla files are gone?

we don't officially support hacked images... because now I can't tell what happened to it.

But let's have a go.

can you go to that directory?

Code:

cd /var/www/system

are all files there? like the one it's complaining about? monitor.sh ?

also to check if the disk is full can you type in this command?

Code:

df -h

and copy paste the result here?
The 3.8 gig image should have almost 1 gig free disc space

This is plenty as the logs get zipped and cleaned after a while.

also please make sure you don't use the PI for anything else other than controlling the miner! Using it for other things other than a controller might flood the logs faster than the logrotator can clean.

We have tested the images extensively on multiple devices, on different locations, Hongkong and Europe ( but only with our own hardware) without problems.

It's an unmodified image and when you reboot it works again for a while, It actually got 20GB of space as I gparted it as you suggested it may be the SD card that's the problem I will try it with one I know has worked for over a week

This is the exact same problem I am having. The PI keeps hashing away though but the web server is not accessible. If I reboot, it works for an indefinite period of time, usually a few hours.

The Pi and SD card have worked together with other images. This seems to be a problem with the image, since more than one person is having it.

maardein

sr. member

Activity: 420

Merit: 250

I am trying to set up controla now, but can't extract the zips? I am getting the error 'Unknown format or damaged archive' with both the hashra-mini-controla-v1.4.0.zip as well as the hashra-atlantis-controla-v1.4.0.2.zip.

I am on windows, and using winrar, which has always worked up till now.

dloganbill

full member

Activity: 126

Merit: 100

I ordered a TP-Link 703n and am going to try Lightning Asics firmware. I'm planning to do a side-by-side comparison of LA vs Hashra with 10 miners each.

maxzilla

member

Activity: 112

Merit: 10

Quote from: TyrTheAlpha on April 25, 2014, 08:58:37 AM

Hello All,
Dloganbill: How are displaying your pi on your phone? What software/app are you using for that?

you can open controla in your iphone with your phone browser. It will scale itself to display properly on a mobile device.

TyrTheAlpha

member

Activity: 63

Merit: 10

Hello All,
Dloganbill: How are displaying your pi on your phone? What software/app are you using for that?

Quote from: dloganbill on April 24, 2014, 11:02:39 PM

I seem to have a lot of HW issues with the latest firmware. I've been mining at Nicehash @850. I have two sets of 10 miners running on the hardware Hashra sent. 10 miners per rpi. Any ideas on what is causing this?

maxzilla

member

Activity: 112

Merit: 10

Quote from: bexwhitt on April 25, 2014, 08:45:57 AM

It's an unmodified image and when you reboot it works again for a while, It actually got 20GB of space as I gparted it as you suggested it may be the SD card that's the problem I will try it with one I know has worked for over a week

It looks like it can't access that file for some reason, I have a feeling it might be the card aswell.

bexwhitt

newbie

Activity: 50

Merit: 0

Quote from: maxzilla on April 25, 2014, 04:25:00 AM

Quote from: bexwhitt on April 25, 2014, 03:21:34 AM

1.4 runs for a few hours then you can't connect to the web interface, I get

Forbidden

You don't have permission to access / on this server.

if you look at the raspberry pi output it's doing this

https://www.youtube.com/watch?v=48bALCHC0qI&feature=youtu.be

the miners seem to be still working though

hmm, it looks like the controla files are gone?

we don't officially support hacked images... because now I can't tell what happened to it.

But let's have a go.

can you go to that directory?

Code:

cd /var/www/system

are all files there? like the one it's complaining about? monitor.sh ?

also to check if the disk is full can you type in this command?

Code:

df -h

and copy paste the result here?
The 3.8 gig image should have almost 1 gig free disc space

This is plenty as the logs get zipped and cleaned after a while.

also please make sure you don't use the PI for anything else other than controlling the miner! Using it for other things other than a controller might flood the logs faster than the logrotator can clean.

We have tested the images extensively on multiple devices, on different locations, Hongkong and Europe ( but only with our own hardware) without problems.

It's an unmodified image and when you reboot it works again for a while, It actually got 20GB of space as I gparted it as you suggested it may be the SD card that's the problem I will try it with one I know has worked for over a week

maxzilla

member

Activity: 112

Merit: 10

Quote from: dloganbill on April 25, 2014, 06:56:43 AM

Here is my history from Nicehash
https://www.nicehash.com/?p=miners&a=0&d=1&addr=192wH2yhWVyjiUkzf2qppnATSvmQkuMsxA

They use a different stratum port for sha and nscrypt. If you compare my log to many if the other miners, mine looks very unstable.
https://www.nicehash.com/index.jsp?p=miners

according to the pool stats you should be getting around 5.6 btc/gh/day = 0.0056 btc/mh/day = 0.0364/6.5mh/day (assuming you have 6.5 mh)
Today in 12 hours you got paid 0.028 btc already. so it's inline with the pool stats if not better.

I'm guessing the hw errors are the pool sending you scrypt-n hashes which will give an invalid nonce.
Compared to other miners, yours looks normal in my opinion, around 2% mhs reject rate.
I wouldn't worry about it.

dloganbill

full member

Activity: 126

Merit: 100

Here is my history from Nicehash
https://www.nicehash.com/?p=miners&a=0&d=1&addr=192wH2yhWVyjiUkzf2qppnATSvmQkuMsxA

They use a different stratum port for sha and nscrypt. If you compare my log to many of the other miners, mine looks very unstable.
https://www.nicehash.com/index.jsp?p=miners

maxzilla

member

Activity: 112

Merit: 10

Quote from: dloganbill on April 24, 2014, 11:02:39 PM

I seem to have a lot of HW issues with the latest firmware. I've been mining at Nicehash @850. I have two sets of 10 miners running on the hardware Hashra sent. 10 miners per rpi. Any ideas on what is causing this?

How is it showing up on the pool side? Does your shares go through fine? Are you getting pay outs as expected?

Can you also try another pool? Do you get the same HW? if not then it's pool related. The pool might be mining a different algo coin, like scrypt-n which is not possible with gridseed chips

maxzilla

member

Activity: 112

Merit: 10

Quote from: bexwhitt on April 25, 2014, 03:21:34 AM

1.4 runs for a few hours then you can't connect to the web interface, I get

Forbidden

You don't have permission to access / on this server.

if you look at the raspberry pi output it's doing this

https://www.youtube.com/watch?v=48bALCHC0qI&feature=youtu.be

the miners seem to be still working though

hmm, it looks like the controla files are gone?

we don't officially support hacked images... because now I can't tell what happened to it.

But let's have a go.

can you go to that directory?

Code:

cd /var/www/system

are all files there? like the one it's complaining about? monitor.sh ?

also to check if the disk is full can you type in this command?

Code:

df -h

and copy paste the result here?
The 3.8 gig image should have almost 1 gig free disc space

This is plenty as the logs get zipped and cleaned after a while.

also please make sure you don't use the PI for anything else other than controlling the miner! Using it for other things other than a controller might flood the logs faster than the logrotator can clean.

We have tested the images extensively on multiple devices, on different locations, Hongkong and Europe ( but only with our own hardware) without problems.

bexwhitt

newbie

Activity: 50

Merit: 0

1.4 runs for a few hours then you can't connect to the web interface, I get

Forbidden

You don't have permission to access / on this server.

if you look at the raspberry pi output it's doing this

https://www.youtube.com/watch?v=48bALCHC0qI&feature=youtu.be

the miners seem to be still working though

dloganbill

full member

Activity: 126

Merit: 100

I seem to have a lot of HW issues with the firmware. I've been mining at Nicehash @850. I have two sets of 10 miners running on the hardware Hashra sent. 10 miners per rpi. Any ideas on what is causing this?

KryptoKings

hero member

Activity: 854

Merit: 506

I got my blade as high as 5550 khs at 800 mhz. Seems stable. No HW errors. Nice Job on the software! Thanks.

lajz99

hero member

Activity: 756

Merit: 500

Up and running with a blade but can't push it past 800 Mhz, will you guys be adding more gradual steps up to 850Mhz?

fivejonnyfive

sr. member

Activity: 294

Merit: 250

Not sure if anyone else has had this problem - but after every couple hours I'm unable to log into the web interface.

Curiously I get a "The requested URL /index.php was not found on this server." error message. Seems like something on apache is borked but I can't SSH into the damn thing to do any kind of troubleshooting.

Any thoughts would be appreciated. I'm getting ready to give up and just install raspbian and bfgminer and write a script to autolaunch the miner since the web interface isn't working on this version anyway....

maxzilla

member

Activity: 112

Merit: 10

Quote from: cereal7802 on April 24, 2014, 01:36:23 PM

Quote from: miter_myles on April 24, 2014, 12:44:41 PM

I'm not in front of mine right now..

But - when attempting to login it would loop back to the login screen. If I entered the wrong password it would spit back invalid creds (red font)... but trying to retry the good creds, it would almost seem like it would think for a second then just redirect to the login screen, blank out the creds..

I haven't tried a hard reboot (as the hashing was going just fine looking at pool stats).. I figured once my auto 6hr reboot occurred I would be able to get in.. but still haven't been able to.

I may be able to dig more up later if I have time.

Current config - was 18 miners on the PI split into two powered hubs, scrypt mode. I have 3 other PIs setup with the older version still (havent had time to reflash all card).. no issues logging in to the older ones (going thru same router/hubs/etc).

sounds like what mine did. turned out the disk was full because hashra image logs a ton of stuff and doesn;t seem to cleanup the logs. perhaps it is a result of an older image being updated, but i have a script that cleans the log files now every so often.

hmm plausible... although logrotate is installed on the image..that should clean the logs every now and then.
you can also mount the image on a linux box and use gparted to resize the partition to 8 or 16 gigs. The 3.8 gig image should have around 1 gig free space.

lajz99

hero member

Activity: 756

Merit: 500

I have been told HASHRA BLADE will be available on the support page "in just a few hours"!!!! Grin

cereal7802

newbie

Activity: 57

Merit: 0

Quote from: miter_myles on April 24, 2014, 12:44:41 PM

I'm not in front of mine right now..

But - when attempting to login it would loop back to the login screen. If I entered the wrong password it would spit back invalid creds (red font)... but trying to retry the good creds, it would almost seem like it would think for a second then just redirect to the login screen, blank out the creds..

I haven't tried a hard reboot (as the hashing was going just fine looking at pool stats).. I figured once my auto 6hr reboot occurred I would be able to get in.. but still haven't been able to.

I may be able to dig more up later if I have time.

Current config - was 18 miners on the PI split into two powered hubs, scrypt mode. I have 3 other PIs setup with the older version still (havent had time to reflash all card).. no issues logging in to the older ones (going thru same router/hubs/etc).

sounds like what mine did. turned out the disk was full because hashra image logs a ton of stuff and doesn;t seem to cleanup the logs. perhaps it is a result of an older image being updated, but i have a script that cleans the log files now every so often.

Topic: HASHRA CONTROLA on Raspberry Pi for Gridseed - page 7. (Read 39630 times)