Tyan S8030GM2NE

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

anything1233

Just a Member
Dec 29, 2020
41
18
8
Does anyone know if this board works with 12V-only PSU? I bought the board expecting to use a regular server PSU, but seems like it requires 5V and 3.3V the way the consumer boards do.

The manual mentions that it supports EPS 12V PSUs. I couldn’t really find what a EPS 12V PSU is… There are consumer PSUs that have EPS 12V connector, so is that it? Or does it mean what I think it does – I can use 12V only PSU?

Has anyone tried connecting just 2 8-pin and not connecting the main 24-pin connector to see if it works?
 
Last edited:

redeamon

Active Member
Jun 10, 2018
291
207
43
As far as I know, all motherboards require 3.3/5. If the power supply is 12v only (most server p/s)- either there's a PDU before the m/b or the m/b has those components built in.


EPS is extra power just for the CPU.

If you're really pressed you might be able to find some buck convertor boards that support 12v->24pin. Something like: https://www.amazon.com/RGEEK-24pin-Input-Output-Switch/dp/B07PPYWGNH
 
Last edited:

RolloZ170

Well-Known Member
Apr 24, 2016
5,140
1,546
113
Has anyone tried connecting just 2 8-pin and not connecting the main 24-pin connector to see if it works?
will not work. can damage the board, i think not.
The manual mentions that the board supports EPS 12V PSUs
EPS Specification
Entry-Level Power Supply Specification (EPS), which is a derivative of the ATX specification, was released for high-end PCs and entry-level servers. It was developed and released by the Server System Infrastructure forum. In order for a PSU to meet the EPS specification, it must have a 24-pin motherboard connector and an EPS (eight-pin) connector. It must also have one four-pin 12V connector if its capacity is 700 to 800W, or two four-pin 12V connectors for capacities of over 850W. The latest EPS specification is 2.92

 

anything1233

Just a Member
Dec 29, 2020
41
18
8
Oh ok, it makes sense now. Thanks!

I was looking at pico PSUs, but the whole point of server PSU was reliability, and I’m not sure how reliable are pico. I know it’s just a buck converter, but still…
 

hmw

Active Member
Apr 29, 2019
570
226
43
Oh ok, it makes sense now. Thanks!

I was looking at pico PSUs, but the whole point of server PSU was reliability, and I’m not sure how reliable are pico. I know it’s just a buck converter, but still…
You can get DC-DC PSUs that have more wattage than the PicoPSU. Like this: HDPLEX 400W HiFi DC-ATX. There are also other PSUs on Alibaba with higher wattage (the X7-ATX-500W), the problem is finding a 19V 20A DC power brick - a lot of the commonly available power bricks top out at 330W: Dell Alienware M18x 330W AC Adapter XM3C3 ADP-330AB B DA330PM111 - Newegg.com.

Consider that the Tyan with a 32 core CPU and 8 channels of RAM takes 300W ~ 350W at load, you will need a 500W PSU to be on the safe side. Hence it's a lot more common for embedded or mATX/Mini-ITX form factors to take 12V or 19-24V e.g. the X11SBA from SuperMicro ...

1671479258977.png

The only non-embedded motherboards that take solely 12V are the ones with the new 12VO standard.
 

Bryce2113

New Member
May 15, 2020
10
8
3
Enabling Remote BIOS

Remember the BIOS page the Tyan IPMI interface has?

View attachment 20427


Does your BIOS look like this even after using the default credentials Administrator/superuser ?

View attachment 20426

Don't worry - it's because the Redfish API the BIOS page depends on getting the default password changed (probably because of CA law SB-327 and that's probably why Dell can show remote BIOS screens). All you need is curl / Python / Postman or anything that can send and receive web requests to the Redfish API

Step 1: We need to get the etag value from the server Redfish API (we will use it in the request header later)
Code:
curl -k \
     --request GET 'https://192.168.0.10/redfish/v1/'
Gives us
Code:
{
  "@odata.context":"/redfish/v1/$metadata#ServiceRoot.ServiceRoot",
  "@odata.etag":"W/\"<some_number>\"",
  "@odata.id":"/redfish/v1/",
  "@odata.type":"#ServiceRoot.v1_4_0.ServiceRoot
...
}
Step 2: In the result above, note down the value for the etag W/"<some_number>" without the <> of course. Now remember the default Administrator:superuser username and password? We will pass it to the Redfish API and see what the problem is.
Code:
curl -k -u 'Administrator:superuser' \
        --request GET 'https://192.168.0.10/redfish/v1/AccountService/Accounts/1' \
        --header 'If-Match: W/"<some_number>"'
Gives us
Code:
{
    "error":
    {
        "@Message.ExtendedInfo":
        [
            {
                "@odata.type": "#Message.v1_0_5.Message",
                "Message": "The password provided for this account must be changed before access is granted.  PATCH the 'Password' property for this account located at the target URI '/redfish/v1/AccountService/Accounts/1' to complete this process.",
                "MessageArgs":
                [
                    "/redfish/v1/AccountService/Accounts/1"
                ],
                "MessageId": "Base.1.0.0.PasswordChangeRequired",
                "RelatedProperties":
                [],
                "Resolution": "Change the password for this account using a PATCH to the 'Password' property at the URI provided.",
                "Severity": "Critical"
            }
        ],
        "code": "Base.1.0.0.PasswordChangeRequired",
        "message": "The password provided for this account must be changed before access is granted.  PATCH the 'Password' property for this account located at the target URI '/redfish/v1/AccountService/Accounts/1' to complete this process."
    }
}
Step 3: So now let's use curl to change the default Redfish account username and password to <new_user> and <new_passwd> (Substitute accordingly)
Code:
curl -k -u 'Administrator:superuser' \
        --request PATCH 'https://192.168.0.10/redfish/v1/AccountService/Accounts/1' \
        --header 'If-Match: W/"<some_number>"' \
        --header 'Content-Type: application/json' \
        --data-raw '{"UserName":"<new_user>", "Password":"<new_passwd>" }'
This should give a HTTP 204 (no content) - but I actually got an error message, the command still worked though. Now when you use the BIOS link to login and use your new username and password, you should be able to see the BIOS settings and even change them

View attachment 20428

Hope this helps ...
I'm having issues getting the BIOS login reset using curl. My odata.etag is a bit different from the example you gave:
1672518222171.png

I get a similar error though later saying it needs to patch the property:
1672518985443.png

I then run different versions of the script you provided and get met with either "URL using bad/illegal format or missing URL" ; failure to resolve host; or "unmatched close brace/bracket" (which I'm guess is from the -data-raw command (chosen password redacted):

1672518514629.png

I'm a relative noob when it comes to curl (I've probably used it 2-3 times in my life but never to parse this much info). Any help/guidance is appreciated!
 

Attachments

Last edited:

hmw

Active Member
Apr 29, 2019
570
226
43
I'm having issues getting the BIOS login reset using curl. My odata.etag is a bit different from the example you gave:

...
I'm a relative noob when it comes to curl (I've probably used it 2-3 times in my life but never to parse this much info). Any help/guidance is appreciated!
Your problem is Windows. The Windows CMD prompt does not support single quotes. One alternative would be to save the payload to be sent into a file

JSON:
{
  "UserName":"<new_user>",
  "Password":"<new_passwd>"
}
let's say you call it json.txt - then the command becomes

Code:
curl -k -u 'Administrator:superuser' \
        --request PATCH 'https://192.168.0.10/redfish/v1/AccountService/Accounts/1' \
        --header 'If-Match: W/"<some_number>"' \
        --header 'Content-Type: application/json' \
        --data-binary @json.txt
The other alternative would be to use WSL on Windows or even a RPi and use a proper bash/zsh shell ...
 

Bryce2113

New Member
May 15, 2020
10
8
3
I'm having issues getting the BIOS login reset using curl. My odata.etag is a bit different from the example you gave:
View attachment 26411

I get a similar error though later saying it needs to patch the property:
View attachment 26414

I then run different versions of the script you provided and get met with either "URL using bad/illegal format or missing URL" ; failure to resolve host; or "unmatched close brace/bracket" (which I'm guess is from the -data-raw command (chosen password redacted):

View attachment 26413

I'm a relative noob when it comes to curl (I've probably used it 2-3 times in my life but never to parse this much info). Any help/guidance is appreciated!
UPDATE: I think I sorted out the URL issues but now it seems the server is expecting a different content type than I'm providing:

1672521283811.png
 

Bryce2113

New Member
May 15, 2020
10
8
3
Your problem is Windows. The Windows CMD prompt does not support single quotes. One alternative would be to save the payload to be sent into a file

JSON:
{
  "UserName":"<new_user>",
  "Password":"<new_passwd>"
}
let's say you call it json.txt - then the command becomes

Code:
curl -k -u 'Administrator:superuser' \
        --request PATCH 'https://192.168.0.10/redfish/v1/AccountService/Accounts/1' \
        --header 'If-Match: W/"<some_number>"' \
        --header 'Content-Type: application/json' \
        --data-binary @json.txt
The other alternative would be to use WSL on Windows or even a RPi and use a proper bash/zsh shell ...

Thanks!!! I ended up installing WSL (Ubuntu) on Windows and that did the trick. Oddly I had to get the W number again as it changed for some reason but otherwise Linux FTW.
 

Bryce2113

New Member
May 15, 2020
10
8
3
Random issue I'm hoping someone else has come across. I've been having instability with the system since I got it up and running. In windows I was getting kernel power codes and figured it was something with the CPU/RAM/OS/PSU/etc. To troubleshoot I installed Unraid last night (which is what the server will ultimately run) and let it sit. This morning I come back to the system locked at the boot screen.

According to the IPMI log, there were a bunch of errant "button_or_switch BUTTON logged a reset button pressed" or a ". . . . logged a power button pressed".
1673022526912.png

The oddity is no one touched the button. And when I go back to the previous crash there is a "button_or_switch BUTTON logged a reset button pressed" error at 4:01AM (I'm a night owl but not that much). Especially odd is that I did not have any FP connectors plugged into the board at this point. . . .

So the question is has anyone ever encountered the FP header on the board going bad? I bought it used and I noticed it came with some hot glue residue on/near the power switch area. Now I'm starting to wonder if the previous owner did that to deal with this issue.

Any thoughts or tips appreciated!
 

hmw

Active Member
Apr 29, 2019
570
226
43
Random issue I'm hoping someone else has come across. I've been having instability with the system since I got it up and running. In windows I was getting kernel power codes and figured it was something with the CPU/RAM/OS/PSU/etc. To troubleshoot I installed Unraid last night (which is what the server will ultimately run) and let it sit. This morning I come back to the system locked at the boot screen.

According to the IPMI log, there were a bunch of errant "button_or_switch BUTTON logged a reset button pressed" or a ". . . . logged a power button pressed".
View attachment 26479

The oddity is no one touched the button. And when I go back to the previous crash there is a "button_or_switch BUTTON logged a reset button pressed" error at 4:01AM (I'm a night owl but not that much). Especially odd is that I did not have any FP connectors plugged into the board at this point. . . .

So the question is has anyone ever encountered the FP header on the board going bad? I bought it used and I noticed it came with some hot glue residue on/near the power switch area. Now I'm starting to wonder if the previous owner did that to deal with this issue.

Any thoughts or tips appreciated!
I installed a Seasonic Prime TX PSU and I had this happen - it went away when I switched back to the original PSU which was a Seasonic Focus PX. What PSU are you using on this motherboard?
 

RolloZ170

Well-Known Member
Apr 24, 2016
5,140
1,546
113
The oddity is no one touched the button
afaik if you click "restart PC" i.e. prompted after update install the OS produces a reset and or power button event.
probably after a critical error the OS initiates a autom. reboot and you get those entries.
 

Bryce2113

New Member
May 15, 2020
10
8
3
afaik if you click "restart PC" i.e. prompted after update install the OS produces a reset and or power button event.
probably after a critical error the OS initiates a autom. reboot and you get those entries.
Perhaps, but there are multiple presses registered, even after the OS was no longer posted, which hard locked the system so it was stuck at the post screen. And Unraid will not auto update restart to install updates.
 

Bryce2113

New Member
May 15, 2020
10
8
3
I installed a Seasonic Prime TX PSU and I had this happen - it went away when I switched back to the original PSU which was a Seasonic Focus PX. What PSU are you using on this motherboard?
Using Supermicro redundant 2U 750W platinum units (PWS-741P-1R) (running through the SM power distro unit the PSUs came with). Not sure if its the PSUs or what. Its odd that for you downgrading to a technically lower class of Seasonic PSU (which all of their stuff is usually solid) would fix the issue.

I've locked/disabled the FrontPanel in the IPMI, so theoretically any button presses show in the log but are ignored (at least it works in my test so far).
 

Bryce2113

New Member
May 15, 2020
10
8
3
Using Supermicro redundant 2U 750W platinum units (PWS-741P-1R) (running through the SM power distro unit the PSUs came with). Not sure if its the PSUs or what. Its odd that for you downgrading to a technically lower class of Seasonic PSU (which all of their stuff is usually solid) would fix the issue.

I've locked/disabled the FrontPanel in the IPMI, so theoretically any button presses show in the log but are ignored (at least it works in my test so far).
Based on this power supply theory, I started experimenting. Since I'm using the SM redundant PSUs, I had their PSMI cable plugged into the Tyan S8030. I figured this would allow me to monitor the PSU vitals and if a PSU dropped the system would stay alive but give me a persistent error beep/noise to alert me.

Well with the PSMI cable plugged in, when I unplugged one of the PSUs the system immediately crashed--seemingly a hard reset but not graceful. If I remove the PSMI cable, when I unplug the PSU, the system stays up and no codes are logged.

So seems there is some credence to the PSU issue. Guessing some instability with the PSUs hit (or something with their UPS delivery) causing the PSMI to trigger. Not sure why the system would lock/crash versus continue running but with an error. But perhaps the SM unit doesn't talk well wth what the Tyan is expecting. Will continue to monitor.
 

RolloZ170

Well-Known Member
Apr 24, 2016
5,140
1,546
113
well the PDB unit can be th issue. there is many electronics (build 3.3/5V i.e.) on it and I2C sensors.