Topic: Avalon ASIC users thread - page 117. (Read 438596 times)

today I have experienced a "bigger problem" for the second time since I have it

the machine was unresponsive on web interface, so I thought it has crashed completely and would need a physical restart, which is very bad for me

however, I tried to connect to SSH and somehow it worked, but extremely slowly, took my a couple of minutes but finally I was able to type reboot command and reboot the machine, which it did and everything is OK again

so, if someone has this totally unresponsive machine, try SSH and be very very patient, it may come up

funny thing is that it did not stop hashing until I reboot it - i.e. even in this strange unresponsive state, it was doing its job, just web interface could not be reached and SSH extremely hardly

since the operation with the whole machine was impossibly slow I could not analyze any logs before reboot, so I have no idea what happened

That is definitely far more than you should be getting. There is a good chance you're submitting heaps of duplicates which may also be a different form of instability that the auto mode can't check for. Try setting a lower maximum speed if you're using auto mode because clearly you're not doing 82GH of useful work.

It seems like I have way more than other Avalon users, at least from what I've seen on Ozcoin. Most of them have only a couple percent invalid shares at most, while I will have 12 or 13% invalid shares. Right now I'm standing at about 94% efficiency on average, that's with 400,000 invalid shares in the less than a week I've been mining there.

This is normal on stratum. Every time the block changes, you get invalids. Some days burn more blocks than others.

I get a REALLY large number of invalid shares with my Avalon.

I'm not sure whats going on. It's not consistent.

One day I'll be mining and I'll get 99+% valid shares and the next day or two it will only be 88% valid shares.

I've tried mining on Ozcoin and 50btc and it seems to be bad on both.

I don't think it is my network because my jalapenos only get <1% invalid shares.

I'm using the dynamically adjusting frequency and it usually settle at about 352 MHz or so and 82 GH/s on average. Of course, according to the pool, I'm only getting 70 GH/s or so due to all the invalid shares.

I have an AC feeding directly into the intake on the Avalon and no temperature gets above 48 or 49 or so (very rarely 50).

What could be wrong?

That's why I have a backup pool and use failover mode.  I even have eloipool running locally, set as the secondary backup, so even in the internet fails completely, it can still mine against that (until it becomes out-of-sync if the internet failure lasts long enough).



If the fans stop running entirely, does that indicate FPGA failure?

My Batch 1 unit does this sometimes.  I now have a script that queries the API, and if it has a zero hashrate for too long, it will call the connected Web Power Switch to cycle the power on the unit, which always brings it back up.  (I do have to have it leave the power off for a full 30 seconds.)

you never know.  the avalon that has given me the most trouble was the last one received about a week and a half ago.

the deletion of the WWAN has so far worked to eliminate the -71 and i haven't seen one for over 3d now.

It is one of the last batch #2 models, so I hope these older problems are not present. I just hope and do not check since I have it in server housing without an easy physical access to the machine, so I would do such check only if there is a serious danger if this is not done. I hope there is no such danger.

Each module has 8 boards connected to a backpanel so in total you have 24 miners with 10 asic chips each one

Easy visualization: https://bitcointalksearch.org/topic/avalon-clone-assembly-service-europe-210186

what will make a Stat1 revert back to Stat0?   a hard reboot?

what do the other 14 represent?

Thats how many asic chip there are per the 24 miner count .. =)

why does ASIC count only read 10?

No such thing as "appropriate" as it's just another hardware error. If your hashrate is okay, it's okay. You can read a few pages ago about how I chose the "appropriate" hardware error target rate of <2%.

Con, what is the appropriate NMW/MWC ratio?

Most 3 module systems running with --avalon-auto mode run perfectly stable.

I have one 4 module test system and that runs into this problem every few hours, the longest I've seen up is 18 hours.

Symptoms are: no mining activity on the pool, the TP-LINK web page displays value 0 for all measurements like temp and fan RPM, the only counter increasing (rapidly) is HW errors. The yellow LED on the back is off so it thinks it mining and it still consumes power (and produces heat) like it's running at 300 MHs

By the way, the 2 module system I have now also ran at a low speed when I set the config to miner count: 16, I had to set it to 24 (even though there are only 16 miner PCBs).

The stats look kind of funny:


There is fantom work being done on non-existing mining PCBs? ;-)

You found it then. At some stage the avalon disconnects due to some USB issue and cgminer automatically hotplugs it. Since it's newly hotplugged it comes up as a new device (ava1 instead of ava0).

Guys, guys, there are no shortcuts. As a student of DE, I need to leave this here and If you have a problem with simple tasks you should take a look. If not, watch as well, its quite entertaining.

http://www.youtube.com/watch?v=eoOUBETTyMI

haha.

last Tuesday i decided to buy a standalone AC from Sears just like someone above had recommended.  it's a pretty big unit 75 lb unit which is all they had.  my avalons are stacked 3 high on a roller rack inside a small open door closet about 6 ft deep 3 feet wide with the fans blowing out into the hallway in my office.  i got the bright idea that i'd put the AC in the back of the closet to provide cold air into the intake fans at the back of the avalons.  note that the roller rack fills up most of the width of the closet and i can just squeeze by thru a space of about 10 inches.

anyways, these AC's produce water into a tray inside the bottom of the unit that needs to be drained at least daily if you're going to run them 24/7.  within a few hours it becomes clear that this is not going to work.  i decide i want to return it.  there was  just not enough space and despite channeling the hot air exhaust ducting out the front of the closet temps don't appear to be any lower than what i was getting with a simple floor fan.  

so, b/c i'm too lazy to pull the roller rack out of the closet b/c of the ethernet, power, server hookups, and most importantly the ongoing hashing (yes, still plugged in and working ), i send my strong male employee into the back of the closet thru that 10" space on the side and tell him to lift the 75lb AC unit over the top of the roller rack with the avalons and hand it to me in the hallway.  so he's struggling to lift the thing, simply b/c of all the wires to navigate around.  he begins to lose control of it as he's passing it over the top.  i quickly grab it and begin to pull it over the avalons when i hear him say "shit" and i hear the splashing of water coming out of the AC.  i quickly pull it entirely out of the closet and set it down expecting to see water all over the top avalon.  all the water missed the top avalon and fell thru that 10" space to the floor.

i learned this; i am one lazy mofo. in my defense, i had totally forgot about the water.

I just had a huge scare I thought I'd share.

The week I was receiving my Avalon's (3), I was moving houses.  At the new house, there was a pre-built shelf in the basement by one of the windows.  I thought, "Great, I'll pop open the window and my Avalon's can vent hot air directly out the window!"  The idea was to pull cool air from the basement and push the hot air out the window.  Great plan.

I went ahead and ran a 240v circuit over to the window and ran Ethernet over.  On the day I was moving the equipment, I quickly pulled down each Avalon from the old house and ran then over and hooked them up.  All was great.  I had three Avalons, two side-by-side and the third stacked on top.  The 240v circuit ran all three boxes at about 10amps and I added a switch so I could hard wire all the boxes.

Over the next week, everything was great.  The boxes were staying cool and the fans were running much lower than at the other house.

Then comes today.  I was taking a nap in the afternoon on the couch.  I awoke to the sound of raindrops and all I saw out the window was gray.  It was a good storm.  I called out to my wife, "I hope you didn't leave any windows open at the old house."  After sitting for another 10 minutes or so, I finally got up from the couch and meandered down to the basement to play some Defiance that I just bought on Steam.

As I came to the bottom of the stairs, the horrible realization dawned on me that my Avalon's were still venting out the window.  I rushed to the window to see a puddle of water covering over half of the top Avalon that was stacked sideways across the other two.  I looked at the boxes and saw three green lights.

I booked it up stairs and grabbed the first towel I could find.  I ran back down and threw the towel on top of the Avalon's and immediately switched off the power to each box (in hindsight I should have done this first).  I yanked the power cords and took each box down and quickly wiped each off with a towel and set them power supply side down.

I looked over at the ledge that the boxes had been sitting on.  There was another puddle of water with more dripping from the top of the window.  It was a mess.

I carefully wiped down all the boxes and after a bit of air drying, I cautiously took a blow dryer to the PSUs and made sure that they were completely dry.  I moved to boxes to the other part of the basement and plugged in the first one.  Green light.  I checked the stats, and it was hashing.  I grabbed the second and did the same.  Green and yellow lights, but it came up and started hashing.  The third box (the one with the puddle of water) is still unplugged.

I now realize that I'm going to have to somewhat rethink my cooling strategy.  I also think that KNC's hosted option is looking more and more attractive.