Turning off machine with red light causes a different machine to turn red.

[gmayra]

Hello everyone!

I have a few racks with S9 antminers. Some racks have 24 miners and other have 18. They are all connected to the internet and hashing, but occasionally one or two machines on each rack will have the red FAULT light on. What I usually do is just shut off the machines, give it 15-20 minutes, and then turn them back on, which normally resolves the issue. But today, every time I turned off a machine, another machine's red light would turn on. I shut off 3 or 4 on each rack and a minute later there are 2 or 3 different machines that now have the red light on. Is there anything I can do to resolve this issue? Does anyone know why this happens? Any help is appreciated.

[Ameador1]

That's odd. Any chance of IP address conflicts of some sort? I mean, you'd think you'd have issues right off the bat, but maybe DHCP leases/renewals and something not updating properly on the DHCP server or miners causing a conflict?

Any chance of change in air flows - maybe some kind of feedback from the machine going off that's causing the next ones in line to get hotter?

Any error's on the first machines indicating what their red light status is related to?

Any error's on the secondary machines indicating what their red light status is related to?

[BitMaxz]

Maybe some of your miner turned to static ip itself that may cause the IP conflict issue.
Could you try to reset them all to factory default? It might solve the issue and make sure your router is set to DHCP mode.

Follow this guide: Three Ways to Restore Factory Settings (R4/S9/T9)

[tim-bc]

If you don't have enough exhaust airflow, turning off one of your machines might let hot air out, which might affect the miners nearby.

Do you know what kind of fatal error you are getting? Try checking the kernel log of the miner (Ctrl + F "Fatal").

[Hockeybum]

Are you running on a 3 phase PDU by chance?  I’m having the same issue. The same psu and asic work fine on a single phase pdu then fail when connected to the 3 phase pdu again.  Plugging in certain miners affect other nearby miners immediately.  Might be a power imbalance on the pdu or heat issue (ambient temps of 105 right now where I’m at as well), I’ll be troubleshooting tomorrow with hopefully lower ambient temps in the 90’s.

[OgNasty]

Any chance of IP address conflicts of some sort?

This was my first thought as well. I’ve had similar issues when running multiple wireless access points.

[Ameador1]

I know it's a bit of a pain as I do it, but I have my router to assign specific IP addresses through DHCP to specific machines based on MAC address. This way I set everything else in the network via DHCP, but keep the miners in a specific range and can manage the IPS at the router instead of setting the addresses on each machine directly. Best of both worlds, but if this is a duplicate IP issue, this could resolve it.