FYI, Intel C2000 family of processors: System Fault may lead to dead system.

Redcoat

MVP
Joined
Feb 18, 2014
Messages
2,924
https://forums.freenas.org/index.ph...-and-poor-customer-service.45762/#post-310340
I have been hit by the same issue (see post here).
So is there a better MB option for HomeMade FreeNas servers to replace the ASRock Mini (C2550D4I) ?

There is no point of just replacing it with the same board if the issue if going to hit you again.

I haven't done that research - though I have the 2750 version in my FreeNAS Mini. I believe the key question you will have to answer is the ability of your case to accept an alternate board choice.

However, I believe that your first move should surely be to make a warranty claim - see here: https://forums.freenas.org/index.ph...-and-poor-customer-service.45762/#post-310340. I would want to contact them and make sure that the replacement board would have the platform-level fix applied and not just be a stock MB supplied on a "fingers-crossed" basis. In that thread you will find an email address for "William" - by all reports he has always been helpful to people so you might try a note to him to understand what's being shipped as warranty replacements. Maybe somebody here will be able to tell us - I suspect that a fix to a board is likely readily identifiable ...
 

fta

Contributor
Joined
Apr 6, 2015
Messages
148
I suspect that a fix to a board is likely readily identifiable ...

I'm getting an advance replacement so I can compare the two boards side by side. I have no idea when I'll actually be getting the replacement board, though. 3 weeks and counting...
 

Redcoat

MVP
Joined
Feb 18, 2014
Messages
2,924
I'm getting an advance replacement so I can compare the two boards side by side. I have no idea when I'll actually be getting the replacement board, though. 3 weeks and counting...

Seems like the right tactic. This is an iXsystems replacement isn't it? 3+ weeks seems like a very loooong time ... (as I think forward to what my experience might be if my Mini goes down). I wonder what they are shipping in new systems today.
 

fta

Contributor
Joined
Apr 6, 2015
Messages
148
Seems like the right tactic. This is an iXsystems replacement isn't it? 3+ weeks seems like a very loooong time ... (as I think forward to what my experience might be if my Mini goes down). I wonder what they are shipping in new systems today.

Yes, it is. Since they couldn't even give me an estimate of when I'd get a new board, I broke down and replaced it with a Supermicro X11SL7-F. When everything in your home runs off your NAS, it's hard to wait an unknown amount of time.
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175

Redcoat

MVP
Joined
Feb 18, 2014
Messages
2,924
New systems include fixed boards.

Yes, I was certain that iXsystems wouldn't be shipping the troubled boards. But the inference from fta's reported experience is that they have no inventory of fixed boards if delivery on a warranty replacement is 3+ weeks/indeterminate. I'm now even more glad that I have a fully capable backup system that I will be able to switch to should my Mini go down.
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Yes, I was certain that iXsystems wouldn't be shipping the troubled boards. But the inference from fta's reported experience is that they have no inventory of fixed boards if delivery on a warranty replacement is 3+ weeks/indeterminate. I'm now even more glad that I have a fully capable backup system that I will be able to switch to should my Mini go down.
Supply chains are complicated things. It might be something as trivial as properly-packaged boards not being available (you can't just throw one unit previously shipped in a box of 20 into a random box and call it a day). Maybe ASRock shipped fewer units than planned because Intel ran out of C2750s at the wrong time.
 

nickt

Contributor
Joined
Feb 27, 2015
Messages
131
Hi all,

I only recently stumbled over this latest announcement about the C2xxx series boards failing due to the LPC clock failing, and the subsequent acknowledgement by iXsystems of the fault affecting the FreeNAS Mini. As one of the lucky people who have also experienced an ASRock C2750D4I board give out due to the BMC issue, I'm feeling just a little frustrated. Now that I've nailed the BMC / watchdog problem, am I next going to lose my replacement board to this latest issue in 18 months time?!?

But on a quick read through of an Anandtech article describing the problem, I can't help but wonder whether these two problems aren't actually the same thing. I speculate that the frequent BMC flash rewrites triggered by the watchdog timer may be communicating using the LPC clock. On the ASRock board, then, perhaps the LPC clock gets smashed by all the BMC activity, and the LPC clock gives out. In the Anandtech article, there is a quote from Intel implying that something is causing higher than expect use of the LPC clock:

"a degradation of a circuit element under high use conditions at a rate higher than Intel’s quality goals after multiple years of service."

Of course, I have no idea what I am talking about, but the probability of getting two very obscure board level design faults on the same board with almost identical symptoms seems very low to me. It also seems possible that, although the problem is at the Intel platform level affecting any brand of board using C2xxx chips, it presents itself more readily on the ASRock board thanks to stupid BMC firmware / watchdog behaviour.

Maybe I am just deluding myself as I try to convince myself that my board won't die in 18 months' time...

It would be nice to hear from iXsystems on this, given that they clearly have a good understanding of the problem and much better access to ASRock technical knowledge than the rest of us.

Interesting to note that ASRock has nothing to say for itself on the problem at this point...
 
Last edited:

Arwen

MVP
Joined
May 17, 2014
Messages
3,600
@nickt
Interesting. Two things:
  1. ASRock may not be able to discuss the issue due to NDA, (Non-Disclosure Agreement), with Intel. Maybe the same for iXsystems.
  2. The original implied BMC flash failure indicated that it exceeded it's write life. But, if access to the chip was un-available, it might appear to be a dead flash chip. So you maybe right.
I miss the old days when you could get schematics and chip specifications for hardware. In a previous life I actually helped design micro-controllers. In which case I could see if the BMC data
flash chip was on the LPC bus.
 

nickt

Contributor
Joined
Feb 27, 2015
Messages
131
Ah dammit. My overly optimistic search for comfort was misplaced. I contacted the venerable William at ASRock, who came back to me straight away and confirmed that the BMC issue and the newly reported LPC clock issue are indeed separate and unrelated issues. Fabulous.

William claims that ASRock are still working with Intel to figure out which boards specifically are affected, but this doesn't seem to gel with iXsystems' implicit acknowledgment that the issue must at least affect all ASRock C2740D4I boards manufactured before February 2017.

William also assured me that ASRock is working with Intel to extend the warranty and that - should my board fail in the future - he'll help me.

Seriously: this sucks...
 

Redcoat

MVP
Joined
Feb 18, 2014
Messages
2,924
Ah dammit. My overly optimistic search for comfort was misplaced. I contacted the venerable William at ASRock, who came back to me straight away and confirmed that the BMC issue and the newly reported LPC clock issue are indeed separate and unrelated issues. Fabulous.

William claims that ASRock are still working with Intel to figure out which boards specifically are affected, but this doesn't seem to gel with iXsystems' implicit acknowledgment that the issue must at least affect all ASRock C2740D4I boards manufactured before February 2017.

William also assured me that ASRock is working with Intel to extend the warranty and that - should my board fail in the future - he'll help me.

Seriously: this sucks...

Yes, it does. Thanks for taking the initiative in flushing this out.
 

millst

Contributor
Joined
Feb 2, 2015
Messages
141
Other side.

-tm
 

Attachments

  • mb_top.jpg
    mb_top.jpg
    1,006.8 KB · Views: 830

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Revision 1.00? They must be using fixed silicon, then.
 

fta

Contributor
Joined
Apr 6, 2015
Messages
148
My original 2750 board was rev 1.01. The new board ix sent me is rev 1.02.
 

Scharbauer

Cadet
Joined
Feb 4, 2017
Messages
7
Hi

I bought a Supermicro A1SRi-2758F board in the beginning of April, when I got the board I saw that it was revision 1.00.

Being a bit paranoid I'm wondering if this is a fixed board or not, is there any way to determine that?
I forgot to bring up this question with the store when I bought it, but I will also contact them and see what they say.
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Hi

I bought a Supermicro A1SRi-2758F board in the beginning of April, when I got the board I saw that it was revision 1.00.

Being a bit paranoid I'm wondering if this is a fixed board or not, is there any way to determine that?
I forgot to bring up this question with the store when I bought it, but I will also contact them and see what they say.
Ask Supermicro support, they'll be able to tell you. What's the manufacture date?
 

Scharbauer

Cadet
Joined
Feb 4, 2017
Messages
7
What's the manufacture date?

That information isn't present anywhere as far as I can see. There's no date on the product box and I can't see anything on the board itself. Maybe it's on the board somewhere but I've missed it?

Edit: Ok, after booting and entering bios it became a bit clearer.
In bios:

Version: 1.1a
Build date: 08/27/2015

It seems most likely that this board is not fixed then?
 
Last edited:

Arwen

MVP
Joined
May 17, 2014
Messages
3,600
@Scharbauer
That likely is the BIOS firmware build date, not the system board's build date.
 
Last edited:
Top