Dell C6100 "BMC FRU header checksum bad"?

[carlding]

Dear all,

I have been lurking around this forum for some time and found STH a very nice place for discussion and knowledge share. Deeply influenced by STH, I finally got myself one Dell C6100 server purchased from lextecny2012 on ebay. According to the BIOS info, t was manufactured in 2009.

The only test I have run so far is booting each node from a thumb drive linux distribution which seemed working. However, I noticed that close to the end of the post stage, all nodes showed the error message "BMC FRU header checksum bad". I tried googling around, but have no idea what it means. Just wondering if any forumer here has met the same issue before. Any suggestion/comment will be appreciated.

 

[peterpan]

I actually saw that yesterday as well on my C6100 from lextecny2012, I emailed them asking what the problem was, no reply so far.

Their support was a joke. I tried to get into BIOS with my USB keyboard but could not. I emailed them and asked if the USB Legacy device was disabled, two of their support people said they used regular USB keyboard for their testing. I bet they never went into BIOS, otherwise they would know the problem.

I googled about the error message, saw one entry in user manual and it said to contact Dell support. I supposed the machine was out of warranty, as the service tag could not be verified on Dell website. Remember saw people here were able to locate their service tag and transfer warranty over.

Btw, the BIOS version on that node giving the error was like 1.02, while other 3 nodes are at 1.62. Wonder if updating BIOS would fix it.

Dear all,

I have been lurking around this forum for some time and found STH a very nice place for discussion and knowledge share. Deeply influenced by STH, I finally got myself one Dell C6100 server purchased from lextecny2012 on ebay. According to the BIOS info, t was manufactured in 2009.

The only test I have run so far is booting each node from a thumb drive linux distribution which seemed working. However, I noticed that close to the end of the post stage, all nodes showed the error message "BMC FRU header checksum bad". I tried googling around, but have no idea what it means. Just wondering if any forumer here has met the same issue before. Any suggestion/comment will be appreciated.

 

[Mrlie]

I supposed the machine was out of warranty, as the service tag could not be verified on Dell website. Remember saw people here were able to locate their service tag and transfer warranty over.

http://forums.servethehome.com/processors-motherboards/1947-feedback-dell-c6100.html, read the very first post.

 

[peterpan]

http://forums.servethehome.com/processors-motherboards/1947-feedback-dell-c6100.html, read the very first post.

Thanks for the link, I learned that after my post here.

 

[carlding]

Hi Peterpan,

it seems that mine have the version of 1.05. Not sure if it is even possible to upgrade the BIOS. When I requested the seller to help upgrade my system, he said that his tech staff tried but was not able to upgrade the BIOS.

 

[Dr_Drache]

Hi Peterpan,

it seems that mine have the version of 1.05. Not sure if it is even possible to upgrade the BIOS. When I requested the seller to help upgrade my system, he said that his tech staff tried but was not able to upgrade the BIOS.

if you want to try, you must force flash in bios, then reflash in windows; I had to do the same to my DCS unit, and they all took 1.70 well, YMMV, and you risk a dead node. up to you.

 

[peterpan]

Hi Peterpan,

it seems that mine have the version of 1.05. Not sure if it is even possible to upgrade the BIOS. When I requested the seller to help upgrade my system, he said that his tech staff tried but was not able to upgrade the BIOS.

I got a reply from their support, they said that error message was not affecting performance of server because of custom nature of the firmware. Kind of weird.

Btw, BIOS version on mine was at 1.05 as well.

 

[carlding]

Hi Dr_Drache,

May I know the build date of your system? Did you upgrade from 1.05 to 1.70? Thanks.

 

[Dr_Drache]

Hi Dr_Drache,

May I know the build date of your system? Did you upgrade from 1.05 to 1.70? Thanks.

sometime in 2009, I forget right now where to check (building some cables) i upgraded one node from 1.05 and 3 nodes from 1.10

 

[carlding]

sometime in 2009, I forget right now where to check (building some cables) i upgraded one node from 1.05 and 3 nodes from 1.10

Would you please share the detailed procedure to do the upgrade? Thank you. I think that could benefit many of us.

I read from here that we need to use socflash on a DOS boot disk. I feel a bit confused that a re-flash under windows is required...

 

[Dr_Drache]

Would you please share the detailed procedure to do the upgrade? Thank you. I think that could benefit many of us.

I read from here that we need to use socflash on a DOS boot disk. I feel a bit confused that a re-flash under windows is required...

acually, reading your link, that is exactly what i did, but i got it from the C6100 thread here, IMO, the flash from windows is double check, it's not REQUIRED, but as I see it, it allows another process to check the checksum of the bios, and gives you a chance to "fix it" ; I AM aware the in 99% of the cases, if it boots to windows it's fine; but he did it for some reasons, and I felt it was worth the extra time (had a windows hdd already, super quick to swap and boot for POM - Peice Of Mind) personally, I do that with quite a few bioses (I edit bioses on most devices i own, for various reasons) always flash twice, it's only saved my butt, ONCE (i think, can never be sure) in 10 years of crazy bios flashes, so, totally up to you how you feel about it.

bit of warning, it worked for me, it worked for him, will it work for you? -I don't have the correct response for that, MAYBE? is the closest we can get. each flash is a $200 replacement (or so) if it bricks... so keep that in mind, to me $200 vs any fixes/stabilty the bioses have in them, is worth it. (please, no one try to correct me with some crazy "the only posted one thing in a change log" type of response, any programmer knows, bugs get fixed without documentation every day, there are always fixes, even if they are so minor maybe a SINGLE person notices)

if you really need me to expand, or talk live, I'm doing my best to "live" in the irc channel

 

[carlding]

acually, reading your link, that is exactly what i did, but i got it from the C6100 thread here, IMO, the flash from windows is double check, it's not REQUIRED, but as I see it, it allows another process to check the checksum of the bios, and gives you a chance to "fix it" ; I AM aware the in 99% of the cases, if it boots to windows it's fine; but he did it for some reasons, and I felt it was worth the extra time (had a windows hdd already, super quick to swap and boot for POM - Peice Of Mind) personally, I do that with quite a few bioses (I edit bioses on most devices i own, for various reasons) always flash twice, it's only saved my butt, ONCE (i think, can never be sure) in 10 years of crazy bios flashes, so, totally up to you how you feel about it.

bit of warning, it worked for me, it worked for him, will it work for you? -I don't have the correct response for that, MAYBE? is the closest we can get. each flash is a $200 replacement (or so) if it bricks... so keep that in mind, to me $200 vs any fixes/stabilty the bioses have in them, is worth it. (please, no one try to correct me with some crazy "the only posted one thing in a change log" type of response, any programmer knows, bugs get fixed without documentation every day, there are always fixes, even if they are so minor maybe a SINGLE person notices)

if you really need me to expand, or talk live, I'm doing my best to "live" in the irc channel

Dr_Drache,

Thank you very much. Have managed to upgrade BMC to 1.28 and BIOS to 1.69 for all four nodes. The problem of "BMC FRU header checksum bad" seemed gone! The upgrade was failed at beginning when I mounted floppy image remotely via IPMI. Later on, I made a windows98 boot disk on USB thumbdrive and solved the problem.

I tried to flash the FCB firmware as well but was not successful because I am having the older PIC16 FCB. The fans stayed at 6000 RPM with no load. Is there any way to reduce this "default" fan speed?

 

[Dr_Drache]

Dr_Drache,

Thank you very much. Have managed to upgrade BMC to 1.28 and BIOS to 1.69 for all four nodes. The problem of "BMC FRU header checksum bad" seemed gone! The upgrade was failed at beginning when I mounted floppy image remotely via IPMI. Later on, I made a windows98 boot disk on USB thumbdrive and solved the problem.

I tried to flash the FCB firmware as well but was not successful because I am having the older PIC16 FCB. The fans stayed at 6000 RPM with no load. Is there any way to reduce this "default" fan speed?

IPMI commandline, I'd have to check but it would be very similar to : (this is for setting error level)
<code> ipmitool -H [BMC IP ADDRESS] -U root -P [password] sensor thresh "FCB FAN4" lcr 0 </code>

i can't do it right now, trying to sort out how to make perfect routed cables, without a cable managerment (or hell, my wish a NEAT-PATCH)
so, all by one device is currently.... not connected.

 

[carlding]

IPMI commandline, I'd have to check but it would be very similar to : (this is for setting error level)
<code> ipmitool -H [BMC IP ADDRESS] -U root -P [password] sensor thresh "FCB FAN4" lcr 0 </code>

I was not able to find the ipmitool command to do the right job. The command you mentioned seems to set the lower critical speed only?