S9K temperature related network failure - logs

DonnovanSA

newbie

Activity: 5

Merit: 0

Quote from: BitMaxz on February 14, 2020, 05:00:54 PM

The log stuck on the temp.

Code:

2020-02-14 09:15:30 temperature.c:1363:process_status_value: Fatal Error: network connection lost!

It might be a temperature sensor.

The temp from PCB is fine it detects properly but after detecting the temp on the chip you got a fatal error then connection lost.

Can you check the miner's configuration and find the low power mode then enable it. then test it again. Let's see if it fixes your issue.

Just in case if still doesn't work check the kernel logs for changes and make a screenshot of your miner's status then post them here.

Hi I had the same problem (Is your PSU on top of the Miner where the control board is or your network cable running for more than a meter next to power cables. i removed mine and my psu standing next to the miner again and that error went away.) hopefully that helps for you as well.

DonnovanSA

newbie

Activity: 5

Merit: 0

Quote from: BitMaxz on February 14, 2020, 05:00:54 PM

The log stuck on the temp.

Code:

2020-02-14 09:15:30 temperature.c:1363:process_status_value: Fatal Error: network connection lost!

It might be a temperature sensor.

The temp from PCB is fine it detects properly but after detecting the temp on the chip you got a fatal error then connection lost.

Can you check the miner's configuration and find the low power mode then enable it. then test it again. Let's see if it fixes your issue.

Just in case if still doesn't work check the kernel logs for changes and make a screenshot of your miner's status then post them here.

Had the same problem with my S9K farm found out that it was my uplink causing that, my Main Elect from the DB was running next to the network cable and caused interference I moved my cable away and ran it from a different side of the farm now I don't get any more of those errors

Hi also noted that my psu was on-top of the miner that also interfered with connection mine has been running now with no network lost

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: thierry4wd on February 26, 2020, 05:12:54 PM

Well as long as you can flash a different firmware while booted from the Sdcard then that makes sense, i however, got the original Sdcard recovery program sent to me directly by bitmain after requesting it, i am not sure why do they "hide" that and not simply upload it to their website, that is plain stupid but nothing out of the usual for bitmain to do.

thierry4wd

sr. member

Activity: 446

Merit: 347

Quote from: mikeywith on February 26, 2020, 03:42:24 PM

Ok, I think I understood, the firmware is loaded from the Sdcard and does not flash itself on the NAND memory, something similar to Braiins OS when it was first released, so what does he do if he wants to upgrade to a different firmware? I don't see this firmware running on an SD fixes any issue, what if he wants to use another custom firmware for an example, please explain more.

Yes, runing mode only, no write on NAND ... but, is unlocked, you allowed for upgrade (by web gui after boot) any firmware you want !

Quote from: sxo13 on February 26, 2020, 04:35:00 PM

Im now not nearby the miners but Im pretty sure there is only the hole on the metal and no sd card slot behind it, or only the jumpers are missing to switch on the booting like from older s9 models Huh

...I may be mistaken but something is off, anyway, will get checked shortly and report here the outcome based on the tips I got.

I'm sur, you have SDCARD slot ! Wink

the port is behind Wink

sxo13

newbie

Activity: 14

Merit: 1

Quote from: mikeywith on February 26, 2020, 02:58:35 PM

Where are you exactly trying to located the SDcard slot? I highly think you are mistaken, there MUST be an Sdcard slot on the right side of the ethernet port, just between the port and the IP report button.

Im now not nearby the miners but Im pretty sure there is only the hole on the metal and no sd card slot behind it, or only the jumpers are missing to switch on the booting like from older s9 models Huh

...I may be mistaken but something is off, anyway, will get checked shortly and report here the outcome based on the tips I got.

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: thierry4wd on February 26, 2020, 03:19:45 PM

So, put the SDCARD (1go - 8go), power on miner, and, not remove this ! the miner booting by this sdcard

Ok, I think I understood, the firmware is loaded from the Sdcard and does not flash itself on the NAND memory, something similar to Braiins OS when it was first released, so what does he do if he wants to upgrade to a different firmware? I don't see this firmware running on an SD fixes any issue, what if he wants to use another custom firmware for an example, please explain more.

thierry4wd

sr. member

Activity: 446

Merit: 347

Sorry, my english is not very good ! lol

old = not take off / not remove

So, put the SDCARD (1go - 8go), power on miner, and, not remove this ! the miner booting by this sdcard

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: thierry4wd on February 24, 2020, 03:23:31 PM

Just, past all file on SDCARD, power off miner, put sdcard, power on ! AND OLD SDCARD on miner for running on this firmware !

I don't understand the last part in bold, what do you mean by an old sdcard?

Quote from: sxo13 on February 25, 2020, 12:16:02 PM

Thanks much, the only problem is that my s9k do not have SD card slots.... Sad

Where are you exactly trying to located the SDcard slot? I highly think you are mistaken, there MUST be an Sdcard slot on the right side of the ethernet port, just between the port and the IP report button.

thierry4wd

sr. member

Activity: 446

Merit: 347

HO ! you are sur ! ? please verify ! on my know, all S9K are SDCARD slot !

https://support.bitmain.com/hc/en-us/articles/360033757513-S17-S17Pro-S9-SE-S9k-Z11-control-board-program-recovery-SD-card-flashing-with-customized-PW-

if not ... i'm curious! send picture

sxo13

newbie

Activity: 14

Merit: 1

Thanks much, the only problem is that my s9k do not have SD card slots.... Sad

thierry4wd

sr. member

Activity: 446

Merit: 347

Just try this

https://drive.google.com/file/d/1EievYtO1mpnX8IHd8LyPYS9PLmtvygRm/view?usp=sharing

Is this a firmware bootable by SDCARD, is only ORIGINAL firmware + SSH unlocked

Just, past all file on SDCARD, power off miner, put sdcard, power on ! AND OLD SDCARD on miner for running on this firmware !

This SDCARD is utile for downgrade or install custom firmware Wink

sxo13

newbie

Activity: 14

Merit: 1

Quote from: thierry4wd on February 21, 2020, 05:26:52 PM

Hum ... your kernel log is very strange for me... i thinks used by custom firmware... is true ?

Quote from: mikeywith on February 22, 2020, 05:53:07 PM

As far as I know, there is no custom firmware for S9k yet,I believe vnish should have theirs ready in a few weeks...

Cool, will keep an eye on it!

Quote from: thierry4wd on February 23, 2020, 05:00:31 AM

Two custom firmware already exist, i have buy only one, and run good ! S9K up to 20TH but very high temp (or need very low temp local) ... I don't know if he wants me to reveal his name ... So i assure this custom firmware is good

Lets hear man, nothing to lose anyway Tongue

thanks.

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: thierry4wd on February 23, 2020, 05:00:31 AM

Two custom firmware already exist, i have buy only one, and run good ! S9K up to 20TH but very high temp (or need very low temp local) ... I don't know if he wants me to reveal his name ... So i assure this custom firmware is good

I don't understand why would a firmware developer want to keep their firmware a secret, I guess it's the exact opposite they would want. Anyway, it would be great if you could review the firmware, I am personally looking for a custom firmware for these piece of trash, I am just hoping that all or at least most of the issues I have with my Antminer S9ks are firmware related and that they would potentially go away with a better/another firmware.

I however have trust issues when it comes to custom firmware, there are only a few sources that I would trust, but a review from someone who has been around long enough like yourself will without a doubt make me a lot less worried.

thierry4wd

sr. member

Activity: 446

Merit: 347

Two custom firmware already exist, i have buy only one, and run good ! S9K up to 20TH but very high temp (or need very low temp local) ... I don't know if he wants me to reveal his name ... So i assure this custom firmware is good

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: thierry4wd on February 21, 2020, 05:26:52 PM

Hum ... your kernel log is very strange for me... i thinks used by custom firmware... is true ?

As far as I know, there is no custom firmware for S9k yet,I believe vnish should have theirs ready in a few weeks.

Quote from: ?? on ??

All stock firmware(january build),

Have you tried the latest one? it was released in December 2019> https://service.bitmain.com/support/download?product=Antminer%20S9k

sxo13

newbie

Activity: 14

Merit: 1

Quote from: mikeywith on February 21, 2020, 05:03:52 PM

[...]

Thanks for the tip, will look into the code to set free the fan setting in the UI.

Yes, ventilation is fine, I have a cold/hot aisle setup, strong fans etc., all other miners behave nicely, even the 6KW bitfury machines work like a charm.

I looked into a few if there is any mechanical damage but no...and they would not have stopped all at more or less the same time. They worked less then 2 months.

Quote from: thierry4wd on February 21, 2020, 05:26:52 PM

[...]

All stock firmware(january build), weird that with the eprom but I assume its just one of the usual chinese manufacturing wonders or do you think it can cause any problems?

thierry4wd

sr. member

Activity: 446

Merit: 347

Hum ... your kernel log is very strange for me... i thinks used by custom firmware... is true ?

Reason : your 3 hashboard is not same eprom

Code:

read chain[0] hardware info:
major type: 0
minor type: 0
chip level: 2
bom version: 0x10
pcb version: 0x39
fixture 8pattern result: L0

read chain[1] hardware info:
major type: 0
minor type: 0
chip level: 2
bom version: 0x10
pcb version: 0x39
fixture 8pattern result: L2

read chain[2] hardware info:
major type: 0
minor type: 0
chip level: 2
bom version: 0x10
pcb version: 0x39
fixture 8pattern result: L0

And, this is not conventional :

Code:

2020-02-14 09:07:35 driver-btm-soc.c:4877:get_working_voltage_from_eeprom: get working vol [ 9.90] from chain[0] eeprom
2020-02-14 09:07:35 power.c:232:set_working_voltage_by_chain: chain[0] working_voltage = 9.90
2020-02-14 09:07:35 driver-btm-soc.c:4877:get_working_voltage_from_eeprom: get working vol [ 9.90] from chain[1] eeprom
2020-02-14 09:07:35 power.c:232:set_working_voltage_by_chain: chain[1] working_voltage = 9.90
2020-02-14 09:07:35 driver-btm-soc.c:4877:get_working_voltage_from_eeprom: get working vol [ 9.90] from chain[2] eeprom
2020-02-14 09:07:35 power.c:232:set_working_voltage_by_chain: chain[2] working_voltage = 9.90
2020-02-14 09:07:35 driver-btm-soc.c:5560:update_highest_voltage: chain 0 hpf working voltage 9.90
2020-02-14 09:07:35 driver-btm-soc.c:5560:update_highest_voltage: chain 1 hpf working voltage 9.90
2020-02-14 09:07:35 driver-btm-soc.c:5560:update_highest_voltage: chain 2 hpf working voltage 9.90

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: ?? on ??

1 - no control over the fans, at least not from the UI

Bitmain are doing a great job limiting the things you can do with your miner, but anyway I have figured out an easy way to temper with fan speed in S9k from the GUI, you can read about here

Quote

2 - ventilation is more then good, tried it also at different ambient temperatures but makes no sense to turn off the other machines, its not just 1-2 of them Grin

Depending on the room size, 1-2 miners can make the place very hot, but since you know the ventilation is "more than good" then there is no need to go the route.

did you inspect the gears for physical damage such as lose heatsinks?

sxo13

newbie

Activity: 14

Merit: 1

Quote from: BitMaxz on February 16, 2020, 04:54:29 PM

[...]

Tried the reset earlier, didnt worked....

I read it but my batch is from around september/october

Contacted support in the meanwhile, lets see if they have a solution, for now they asked only for additional screenshots...will post here the outcome.

Quote from: mikeywith on February 17, 2020, 11:25:24 PM

[...]

1 - no control over the fans, at least not from the UI
2 - ventilation is more then good, tried it also at different ambient temperatures but makes no sense to turn off the other machines, its not just 1-2 of them Grin

mikeywith

legendary

Activity: 2436

Merit: 6643

be constructive or S.T.F.U

Quote from: BitMaxz on February 16, 2020, 04:54:29 PM

Well, if all recommendation doesn't work I think the issue still under their firmware which is a common issue on s9k miner.

Read this "WANRNING! Do Not Buy Antminer S9k"

Well, if only he read the topic before purchasing Cheesy

The stock firmware has a lot of issues in regards to reading the temperature of the chips, it has "supposedly" fixed in the later version/s, however, these S9ks suck by default, my S9ks keep losing a board every now and then, sometimes they reboot a few times a day for really no obvious reason, I have actually got to the point where I don't bother with them anymore, I have talked to Taserz and Patrick (Awesome Miner) about a firmware for the S9k, hoping that their firmware will fix these issues, but the of them did not promise me anything yet.

if you are certain the problem you have is temperature related, you can simply do one of the following or both.

1-Set all fans to 90%.
2-Turn the other S9s off to reduce the room temperature.

if the problem goes away, then you need to improve your ventilation system, if it does not, then try to return them to Bitmain while you still can.

Topic: S9K temperature related network failure - logs (Read 482 times)