Avalon 821 & 841

[blocbuilder]

Hey all,

just an introduction, we run a set up in Canada with a Batch of Avalon 821's and 841's and I have to admit, these are not easy machines to keep up. We are having problems keeping up performance for more than 15 hrs at a time, and I don't think its the heat because when it's 33c outside they run fine, but come 5am (EST) and 22c all 3 blocs which hold 821's start collapsing. We reboot and all goes back to normal and stays there for hours before it happens again.

And we run s9's next to the blocs and they have no issues.

Has anyone got any idea what this could be ?

[NotFuzzyWarm]

For a start, read through the Avalon troubleshooting guide pinned to the top of the Support section. Bet ya find the answers there...

Something must be seriously wrong with the setup because I've found the Avalons to be rock-solid.

[blocbuilder]

For a start, read through the Avalon troubleshooting guide pinned to the top of the Support section. Bet ya find the answers there...

Something must be seriously wrong with the setup because I've found the Avalons to be rock-solid.

I agree, they are solid for us for about 15 hrs until we start getting issues. We have no error codes coming, which is strange so the Troubleshooting guide is of little help.

[NotFuzzyWarm]

Bigger thing in the repair guide deals with the API logs and how to read them -- that gives detailed information as to PSU voltages, Vcore, temps, speeds, AUC dongle temps & current draw, etc. Pull a copy of the logs when they are running right and another copy when they act up then compare the two.

edit: and DO NOT post dupes of the same query in other sections - mods will delete them...
If you insist on dupes at least post them in the right areas eg Mining Support and preferably under that Avalon repair thread so everyone can see the question and solution without searching all over the Forum... Software area sure ain't it...

[blocbuilder]

Bigger thing in the repair guide deals with the API logs and how to read them -- that gives detailed information as to PSU voltages, Vcore, temps, speeds, AUC dongle temps & current draw, etc. Pull a copy of the logs when they are running right and another copy when they act up then compare the two.

edit: and DO NOT post dupes of the same query in other sections - mods will delete them...
If you insist on dupes at least post them in the right areas eg Mining Support and preferably under that Avalon repair thread so everyone can see the question and solution without searching all over the Forum... Software area sure ain't it...

Will do this, thanks

btw, these machines are no more than 3 weeks old.

[mgoz]

If all of them drop at same time it sounds like it could be a network or DNS issue. I and others had issues with some units coming with bad fans. They'd run fine for x amount of time then overheat and shut off and be fine after rebooting until crashing again. Canaan will send replacement fans if that's the issue but you'd need to monitor logs before it crashes. Once it crashes there's nothing in the logs. Prior to crashing, my bad fan would show spinning at 100% and 0RPM and then overheat to 150C before shutting off.

[Steamtyme]

That is weird. Can you give a little more info?

Like what each of your blocks consist of.

The circuits they are running on.

Is it always 15 hours to failure, or is does it always happen at 5am?

What are you using to power the controllers?

[blocbuilder]

If all of them drop at same time it sounds like it could be a network or DNS issue. I and others had issues with some units coming with bad fans. They'd run fine for x amount of time then overheat and shut off and be fine after rebooting until crashing again. Canaan will send replacement fans if that's the issue but you'd need to monitor logs before it crashes. Once it crashes there's nothing in the logs. Prior to crashing, my bad fan would show spinning at 100% and 0RPM and then overheat to 150C before shutting off.

Yes, we thought it was a network issue however, the s9's we have don't react at all to the collapse in the Avalon blocs. The overheating, shutoff and crashing again is what we are dealing with today however, we don't have the type of temps that result in a shut off, nor do we have fan issues which are showing up so this is all very strange.

[blocbuilder]

That is weird. Can you give a little more info?

Like what each of your blocks consist of.

The circuits they are running on.

Is it always 15 hours to failure, or is does it always happen at 5am?

What are you using to power the controllers?

The decline usually begins at 5-6am before shut off at 9am... if you look at the data, there seems to be a pattern connected to heat build up, but still a little vague.

Bloc 1,2,3 have 20 x Avalon 821's each. 60 in total.

I wonder whether the Power plugs we ordered in for the Controllers are incorrect. Will check this now.

[NotFuzzyWarm]

I wonder whether the Power plugs we ordered in for the Controllers are incorrect. Will check this now.

If you mean the RasPi PSU wall warts, they need to be rated for at least 2.5A.

When you say 'reboot' are you soft booting (not cycling the power off/on) just the miners, restarting cgminer, or soft booting the RasPi?

[blocbuilder]

If you mean the RasPi PSU wall warts, they need to be rated for at least 2.5A.

When you say 'reboot' are you soft booting (not cycling the power off/on) just the miners, restarting cgminer, or soft booting the RasPi?

yes, we soft reboot the Miners and the cgminer. we had our first collapse this morning, rebooted, then another collapse but now they seem to be holding up. if it was heat, they would be going down every 20 min as we are at peak temp.

[Steamtyme]

Just a thought, I can't remember when/where, but someone was once talking about running into problems when they ran the 20 miners off 1 controller.

Not sure if that is playing a part with your setup, but it may be worth grabbing an extra controller and running 15 per.

When you mentioned temps, what are you doing to remove the warm exhaust? Is there a chance the warm air is short-circuiting around to the intake of the miners? Just putting ideas out there.

Also not sure if I missed it before, but do all the miners in each block go down at the same time?

[rifleman74]

Just a thought, I can't remember when/where, but someone was once talking about running into problems when they ran the 20 miners off 1 controller.

Not sure if that is playing a part with your setup, but it may be worth grabbing an extra controller and running 15 per.

When you mentioned temps, what are you doing to remove the warm exhaust? Is there a chance the warm air is short-circuiting around to the intake of the miners? Just putting ideas out there.

Also not sure if I missed it before, but do all the miners in each block go down at the same time?

This, buy another Rpi and see if the problem continues.