X11DPi-nT BMC crashing

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

gtech1

Member
May 27, 2019
112
8
18
SM replace not with board has VRM bug.
you get rev. 2.0x but not brand new.
I understand that. You're saying that potentially SM used an older board, replaced VRM from rev 1.21 board with new vendor VRM and turned it into a 'new' 2.01A board, correct ? Even so, it's pretty weird that it shows the same behavior as a 1.21 board, unless ... I dunno, this is getting too crazy.
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
Even so, it's pretty weird that it shows the same behavior as a 1.21 board,
??? VRMs burned on yours ? thought you have BMC issues.
maybe you misunderstood something.
VRM bug is one thing,
worn out BMC flash is another, can happen on all boards over time.
but every brand new board can die within weeks, SM can not run them for years to find out any error.
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
well, SM checked the serial # and they say it's: PCB - 2.01A but that it may not have the latest 'ECO' updates
so supermicro confirms that this board not came lately out of the factory(brand new board)
if so it should have all ECO modifications.
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
Also, good guess about the BIOS time but that's not it. There's clearly a different login mechanism and IP than mine
replace battery = time reset.
reset BMC: wrong time in LOG.
set time in BIOS or OS = next LOG entry is correct time.
you can login fine to IPMI ? that's BMC, so it works correct this time.
re register USB devices can caused by many other things, e.g. unstable PSU/PDB, overheating.
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
You're saying that potentially SM used an older board, replaced VRM from rev 1.21 board with new vendor VRM and turned it into a 'new' 2.01A board, correct ?
no, the 1.21 board is thrown in the trash but you don't get a brand new board (with full warranty) as replacement.
 

gtech1

Member
May 27, 2019
112
8
18
replace battery = time reset.
reset BMC: wrong time in LOG.
set time in BIOS or OS = next LOG entry is correct time.
you can login fine to IPMI ? that's BMC, so it works correct this time.
re register USB devices can caused by many other things, e.g. unstable PSU/PDB, overheating.
The bmc also threw a chassis intrusion warning, out of the blue, at the same time the usb devices re-registered. How to explain that ? It sounds like the bmc crashing to me
 

gtech1

Member
May 27, 2019
112
8
18
no, the 1.21 board is thrown in the trash but you don't get a brand new board (with full warranty) as replacement.
Ok, but I never had a 1.21 board. This was supposed to be brand new , not refurbished upgraded, so with all eco updates too.

What should I do here ?
 

EvoDyn

New Member
Jun 23, 2021
18
12
3
no, the 1.21 board is thrown in the trash but you don't get a brand new board (with full warranty) as replacement.
Is this just for X11DPi? or are all X11 DP boards affected? Asking as someone who has a X11DPH-T.
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
The bmc also threw a chassis intrusion warning, out of the blue, at the same time the usb devices re-registered. How to explain that ? It sounds like the bmc crashing to me
if BMC crash, how does it create a LOG ?
how to login IPMI with crashed BMC ?
 

gtech1

Member
May 27, 2019
112
8
18
if BMC crash, how does it create a LOG ?
how to login IPMI with crashed BMC ?
It didn't create a log, it just restarted. Chassis intrusion is default setting to alert so it reverted to that.

How else can you explain the chassis intrusion + usb devices re-registering + onboard nics flapping ?
 

RolloZ170

Well-Known Member
Apr 24, 2016
6,716
2,075
113
how to ensure that beyond purchasing from authorized reseller ? I mean, this board was supposed to be new as well
talk to the seller and make clear your want new board.
but as i sayd before, a brand new board can have issues too. you are not 100% safe, 1 of 1000 is maybe bad.
 
  • Like
Reactions: gtech1

gtech1

Member
May 27, 2019
112
8
18
talk to the seller and make clear your want new board.
but as i sayd before, a brand new board can have issues too. you are not 100% safe, 1 of 1000 is maybe bad.
Thank you for your time and advice on this thread. One last general question. I have been a purchaser of SM hardware for 10+ years, never had any issues. Since the AI "bubble" happened and SM has been shipping a lot of hardware to Nvidia, Meta, Amazon, etc. Did you notice a decrease in quality ? We have several instances of dubious quality and huge delivery delays that we never had before. I can honestly say that we still have older generation SM hardware that runs better than the new stuff.
 

gtech1

Member
May 27, 2019
112
8
18
fair enough. maybe one last question. This is a CPU they installed in a brand new server they shipped to us. From what you know, can new cpus look like this ?

1732289937672.jpeg