Kernel space clean work failed

Problem

Under various conditions, a miner running Bitmain firmware (stock) will go through a partial boot process and then spam the logs with an error similar to the following.

2024-04-24 15:43:08 notice kernel space clean work failed.

Solution

ℹ️

This is a Bitmain log error, but also something we have seen in the field.

If you look prior to the messages stating "kernel space clean work failed", closer to the beginning of the boot process, you may see a PSU error.

2000-01-01 00:00:16 ==========================capability end============================

2000-01-01 00:00:16 chain num = 3
2000-01-01 00:00:16 skip loading levels for now

2000-01-01 00:00:16 load chain 0 eeprom data

2024-04-24 14:58:51 load chain 1 eeprom data

2024-04-24 14:58:52 load chain 2 eeprom data

2024-04-24 14:58:52 i2c_sim_init start

2024-04-24 14:58:52 i2c_sim_init end

2024-04-24 14:58:56 get power version failed

2024-04-24 14:58:56 check power version failed, use v2 protocal to try it again

2024-04-24 14:58:56 IIC_CheckAck:281 ack error

2024-04-24 14:58:56 IIC_CheckAck:281 ack error

2024-04-24 14:58:56 IIC_CheckAck:281 ack error

2024-04-24 14:58:56 IIC_CheckAck:281 ack error

2024-04-24 14:58:56 IIC_SendData checkack 3 times,send data checkack error

This error indicates an issue with the PSU. Standard Bitmain support sites state that a PSU replacement is likely the fix, however we have seen this happen due to a miner bring plugged into a metering device at the source of power. Once the metering device was removed from the power path, the issue disappeared and allowed the machine to boot and hash.

Does LuxOS support mixed hashboard models Hashboard displayed as A BHB42XXX