I2C4 Problem in my Fujitsu server

[SHOT]

Hi, I'm currently running Windows Server 2019 in my RX300 S5 and everything worked fine until one day I start getting these messages from the iRMC Event log

 
and then the fans start spinning up for a while and then down again.

This happens 2/3 times a day but everything works fine in the server. Do you know how I can fix this?

Ps: I don't have the LocalView Display

 

[Phas3L0ck]

Start with the obvious things; reset the BIOS, replace the CMOS battery, check to see if all the fans are running, replace RAM modules (if possible), and definitely check the power supply to make sure both modules are functional.

[Phas3L0ck]

Hey I found your post on Fujitsu's support site! https://forum.ts.fujitsu.com/forum/viewtopic.php?p=155273

Heh, funny how small the net is.

 

Anyhow, over on this page it says to update the firmware (if you can even find it anymore)

Translated: The problem was the firmware from the local service display. there is an update for the server. PIC Flash Firmware - LVP Flash Software for bootable USB stick

[SHOT]

23 minutes ago, Phas3L0ck said:

Start with the obvious things; reset the BIOS, replace the CMOS battery, check to see if all the fans are running, replace RAM modules (if possible), and definitely check the power supply to make sure both modules are functional.

Will try this thanks, I will let you know

 

19 minutes ago, Phas3L0ck said:

Hey I found your post on Fujitsu's support site! https://forum.ts.fujitsu.com/forum/viewtopic.php?p=155273

Heh, funny how small the net is.

 

Anyhow, over on this page it says to update the firmware (if you can even find it anymore)

Translated: The problem was the firmware from the local service display. there is an update for the server. PIC Flash Firmware - LVP Flash Software for bootable USB stick

haha yes, I thought that the official forum was worth a chance. Anyway I'm already running the latest firmware and I'm not using the local display.

[Phas3L0ck]

5 minutes ago, SHOT said:

Will try this thanks, I will let you know

 

haha yes, I thought that the official forum was worth a chance. Anyway I'm already running the latest firmware and I'm not using the local display.

Just to be clear, you updated the control firmware AND the BIOS, right?

I had a problem for a short time in which the system would do strange things if you updated one part but not the other.

[SHOT]

10 hours ago, Phas3L0ck said:

Just to be clear, you updated the control firmware AND the BIOS, right?

I had a problem for a short time in which the system would do strange things if you updated one part but not the other.

Wait, maybe I've miss understood "firmware"... Do you mean the firmware for iRMC, the remote management controller right? If you didn't mean that then no, I've just checked irmc because for now I didn't want to touch the bios ecc since everything was working fine..

[Phas3L0ck]

1 hour ago, SHOT said:

Wait, maybe I've miss understood "firmware"... Do you mean the firmware for iRMC, the remote management controller right? If you didn't mean that then no, I've just checked irmc because for now I didn't want to touch the bios ecc since everything was working fine..

That could be your problem. Double check available updates for iRMC and BIOS to make sure the hardware is getting the right signals.

And you never know, maybe a full update will open more features that were released later (not on all systems, but there are things that get added to version upgrades)

[SHOT]

16 hours ago, Phas3L0ck said:

That could be your problem. Double check available updates for iRMC and BIOS to make sure the hardware is getting the right signals.

And you never know, maybe a full update will open more features that were released later (not on all systems, but there are things that get added to version upgrades)

I've just checked.. the bios is up to date with the latest version released in 2013 

Maybe the reason is windows 10. Is it possible that the OS is interfering with the I2C bus? I should install Linux but I would really like to avoid that for time sake

[Phas3L0ck]

The problem is neither software nor hardware, it's firmware. If the error is only showing in iRMC event log it could be isolated to the management chipset in relation to how it interprets faults and other signals. You said you don't have the LocalView display? that could be a part of it. Maybe this is more of an internal message that happens when the display thing isn't attached.  No way to be sure since I haven't used any systems from Fujitsu, but it's possible.

 

Anyhow, as long as system operation isn't affected, just ignore it and try to go on with it as is.  Older servers have strange issues-- desktops of that generation were bad enough, but for problems like this, sometimes there's just no known explanation.

[SHOT]

14 hours ago, Phas3L0ck said:

The problem is neither software nor hardware, it's firmware. If the error is only showing in iRMC event log it could be isolated to the management chipset in relation to how it interprets faults and other signals. You said you don't have the LocalView display? that could be a part of it. Maybe this is more of an internal message that happens when the display thing isn't attached.  No way to be sure since I haven't used any systems from Fujitsu, but it's possible.

 

Anyhow, as long as system operation isn't affected, just ignore it and try to go on with it as is.  Older servers have strange issues-- desktops of that generation were bad enough, but for problems like this, sometimes there's just no known explanation.

I don't hink that the localview display is the problem because it worked fine before!
Also the bios is getting this log messages: https://imgur.com/a/aAYSuHL

 

 

 

 

[Phas3L0ck]

Not sure if it matters at this point, but did you check all the fans to see if they are all running and properly connected?

I just read a side note somewhere that PWM control goes through I2C. Could be one of those weird fan issues.

[SHOT]

I can't delete the message

[Phas3L0ck]

On 1/24/2021 at 5:36 AM, SHOT said:

I can't delete the message

And I'm guessing you've already tried clearing the BIOS with no luck, right?

Not sure what else it could be then.

[SHOT]

On 1/25/2021 at 11:44 PM, Phas3L0ck said:

And I'm guessing you've already tried clearing the BIOS with no luck, right?

Not sure what else it could be then.

Yes, I've done that..

Also, now I get  and I don't know why. The server reboots itself and I need press F2 (bios startup) to continue the boot 

[Phas3L0ck]

12 minutes ago, SHOT said:

Yes, I've done that..

Also, now I get  and I don't know why. The server reboots itself and I need press F2 (bios startup) to continue the boot 

Okay, you're totally gonna hate me for this, but if that other error has stopped showing, there are a few things to do that might clear these errors.

1. physically switch the CPUs in their sockets (sounds lame, but has an effect on older boards)

2. Find a BIOS update image, even if it is the same version you currently have, re-flash it again and then replace the CMOS battery once more - then clear BIOS and reboot (I recently did this to fix a sensor problem on my own server, which is a newer system)

3. if you have several RAM modules, switch them around-- I don't know why, but different combinations have varying effects on POST errors and system functionality **who even built this thing?!?**

4. Proceed to power cycle the system several times, all the while slamming your head against a nearby wall with the nagging feeling that it might not make a difference  *that's a joke, BTW. (sarcastic but very serious) 2 bad I can't express emotion via text.

 

Is that the only set of errors on screen, or is the I2C4 code still showing?

[SHOT]

11 minutes ago, Phas3L0ck said:

Okay, you're totally gonna hate me for this, but if that other error has stopped showing, there are a few things to do that might clear these errors.

1. physically switch the CPUs in their sockets (sounds lame, but has an effect on older boards)

2. Find a BIOS update image, even if it is the same version you currently have, re-flash it again and then replace the CMOS battery once more - then clear BIOS and reboot (I recently did this to fix a sensor problem on my own server, which is a newer system)

3. if you have several RAM modules, switch them around-- I don't know why, but different combinations have varying effects on POST errors and system functionality **who even built this thing?!?**

4. Proceed to power cycle the system several times, all the while slamming your head against a nearby wall with the nagging feeling that it might not make a difference  *that's a joke, BTW. (sarcastic but very serious) 2 bad I can't express emotion via text.

 

Is that the only set of errors on screen, or is the I2C4 code still showing?

hahahaha btw since the IERR errors appears, the I2C4 erros are not showing. I will try your suggestions.. (will also try the last one)

[Phas3L0ck]

Just now, SHOT said:

hahahaha btw since the IERR errors appears, the I2C4 erros are not showing. I will try your suggestions.. (will also try the last one)

LOL Very glad 2 hear. BIOS problems can be one screwy thing. Can't even begin to tell ya the stories of fixing other systems. Desktops are bad enough, but servers... WTF.

Hope I added some fun to your day (or night)

[SHOT]

1 hour ago, Phas3L0ck said:

Okay, you're totally gonna hate me for this, but if that other error has stopped showing, there are a few things to do that might clear these errors.

1. physically switch the CPUs in their sockets (sounds lame, but has an effect on older boards)

2. Find a BIOS update image, even if it is the same version you currently have, re-flash it again and then replace the CMOS battery once more - then clear BIOS and reboot (I recently did this to fix a sensor problem on my own server, which is a newer system)

3. if you have several RAM modules, switch them around-- I don't know why, but different combinations have varying effects on POST errors and system functionality **who even built this thing?!?**

4. Proceed to power cycle the system several times, all the while slamming your head against a nearby wall with the nagging feeling that it might not make a difference  *that's a joke, BTW. (sarcastic but very serious) 2 bad I can't express emotion via text.

 

Is that the only set of errors on screen, or is the I2C4 code still showing?

Ok now its back

Do I still try to follow your procedure?

[Phas3L0ck]

14 minutes ago, SHOT said:

Ok now its back

Do I still try to follow your procedure?

I'll just go for broke; if you haven't reflashed the BIOS with the same image, do that and see what happens.  If you already did the reflash, try and find an older version of your system BIOS and flash the old image to the chip to replace the current one.

Literally exactly what I had to do a few weeks ago for my sensor problem was to flash an older BIOS version. If possible, get an older BIOS update and use it.

If that doesn't work, nothing will.

[SHOT]

On 1/27/2021 at 11:27 AM, Phas3L0ck said:

I'll just go for broke; if you haven't reflashed the BIOS with the same image, do that and see what happens.  If you already did the reflash, try and find an older version of your system BIOS and flash the old image to the chip to replace the current one.

Literally exactly what I had to do a few weeks ago for my sensor problem was to flash an older BIOS version. If possible, get an older BIOS update and use it.

If that doesn't work, nothing will.

If the BIOS update does not fix can you tell if replacing the motherboard will fix this issue?

[Phas3L0ck]

11 minutes ago, SHOT said:

If the BIOS update does not fix can you tell if replacing the motherboard will fix this issue?

It would rule devices on the board in or out, (CPU, RAM, and iRMC) but that can be prohibitively expensive.

If your board works and runs the OS, there's no point.

 

The problem could still be a loose connection somewhere, or simply a jumper not set right.

[SHOT]

10 hours ago, Phas3L0ck said:

It would rule devices on the board in or out, (CPU, RAM, and iRMC) but that can be prohibitively expensive.

If your board works and runs the OS, there's no point.

 

The problem could still be a loose connection somewhere, or simply a jumper not set right.

I'm asking because I've found one used on eBay for 40€. 

I've opened the entire server and checked every cable, everything is plugged correctly and I can't see any jumper. I really don't know what else can be.

[Phas3L0ck]

1 hour ago, SHOT said:

I'm asking because I've found one used on eBay for 40€. 

I've opened the entire server and checked every cable, everything is plugged correctly and I can't see any jumper. I really don't know what else can be.

If you think it's worth it, go for it.

But I do have one more question; are the fans sitting on a backplane that allows them to be hot-swapped, or are they just connected straight to the board?  **I didn't think of that before because such a feature is very rare outside of older or proprietary systems.

[SHOT]

5 hours ago, Phas3L0ck said:

If you think it's worth it, go for it.

But I do have one more question; are the fans sitting on a backplane that allows them to be hot-swapped, or are they just connected straight to the board?  **I didn't think of that before because such a feature is very rare outside of older or proprietary systems.

The fans are pluged into another board that sits near the mother board.

After looking in internet I may have found that the I2C bus is a cable that runs from the motherboard to the fan backplane and to the hdd backplane. The connector to the motherboard is the same for both of them but some cables from the connector go to the fans and others to the hdd backplate.

[Phas3L0ck]

6 hours ago, SHOT said:

The fans are pluged into another board that sits near the mother board.

After looking in internet I may have found that the I2C bus is a cable that runs from the motherboard to the fan backplane and to the hdd backplane. The connector to the motherboard is the same for both of them but some cables from the connector go to the fans and others to the hdd backplate.

That could be significant. Told ya it has something to do with the fans.

Have no idea why they'd refer to it as an I2C error though.

Keep digging and see if you can find specs and/or a pinout for the cable(s) and the headers that connect, and maybe fixing that will start to solve the errors.