Register for the iXsystems Community to get an ad-free experience and exclusive discounts in our eBay Store.

problems with AsRock boards

Status
Not open for further replies.

Jason Fayre

Junior Member
Joined
Apr 11, 2015
Messages
18
Hello all,
I've recently been having a lot of trouble with my Freenas system. My original board which was a Asrock c2550d4i, failed after a year with what Asrock says was a faulty sensor on one of the power rails.
Asrock sent me another board, which was a c2750d4i. I ran this board for about a week, then started having trouble. The system would start rebooting itself every 5-10 minutes. After I discovered this, I shut down the system and attempted to cold boot it. The system would not come up at all. If I cut the power from the PSU switch on the back and then turned it back on, the fans and HDDs would start spinning up for about 3 seconds, then immediately power down. Nothing I did would get the system to power up.
I could access the BMC, and noticed several mentions of the CPU going into high temperature.
Asrock is now sending me another board, which will be the c2550 that they should have replaced mine with in the first place.
Has anyone seen behavior like this on these boards? My system has now been out of service since the middle of June.
My case is a silverstone ds380.




c
 

nojohnny101

Neophyte Sage
Joined
Dec 3, 2015
Messages
1,474
i have the same setup as you, a c2550d4i in a silverstone ds380. sounds like you are having a string of bad luck. is your freenas sitting behind a UPS?

sorry to hear about that though. are you confident you components are sound? you bought new when you built the machine? post your specs like PSU.
 

Thomas Bardin

Newbie
Joined
Aug 9, 2016
Messages
1
Exactly the same problem for me yesterday evening !

the Mobo is taken back on RMA this day....



Hello all,
I've recently been having a lot of trouble with my Freenas system. My original board which was a Asrock c2550d4i, failed after a year with what Asrock says was a faulty sensor on one of the power rails.
Asrock sent me another board, which was a c2750d4i. I ran this board for about a week, then started having trouble. The system would start rebooting itself every 5-10 minutes. After I discovered this, I shut down the system and attempted to cold boot it. The system would not come up at all. If I cut the power from the PSU switch on the back and then turned it back on, the fans and HDDs would start spinning up for about 3 seconds, then immediately power down. Nothing I did would get the system to power up.
I could access the BMC, and noticed several mentions of the CPU going into high temperature.
Asrock is now sending me another board, which will be the c2550 that they should have replaced mine with in the first place.
Has anyone seen behavior like this on these boards? My system has now been out of service since the middle of June.
My case is a silverstone ds380.




c
 

David L

Newbie
Joined
Aug 9, 2016
Messages
1
My c2550d4i also died a couple of days ago, connecting to BMC works but the system will not POST.
Have swapped PSU and removed memory sticks one at a time without any luck.
Since the board is 1 year and 2 months old the dealer will not RMA, have written to AsRock support to see if they are willing to resolve this (haven't heard back in 2 days so i'm guessing no).
 

Pitfrr

Neophyte Sage
Joined
Feb 10, 2014
Messages
1,012
I have a C2750d4i since 10 months and I'm using it on my test set-up with a picoPSU.
It works quite well so far. I just noticed that the board gets quite hot (I mean more than I would have naively expected), mainly the CPU and the GPU, therefore I added a fan to cool it down.
I didn't try to connect to BMC yet... I should try once though. ;-)
 

Jason Fayre

Junior Member
Joined
Apr 11, 2015
Messages
18
I found the following thread on Reddit:
https://www.reddit.com/r/homelab/comments/4ufa8i/c2750d4i_wont_power_up/
This is very concerning. Can anyone comment on this? I'm wondering if this is new behavior in Freenas 9.10? My problem and and the one mentioned in this thread is identical, even the 15 day timing.
I'm about to get a replacement asrock board, but don't want to use this new board if I'm going to have the same issue.
 
Last edited by a moderator:

DaveY

Member
Joined
Dec 1, 2014
Messages
141
Wow, OP post might as well have been mine. I ended up having to RMA the board. Tech support told me the exact same thing; power rail problems. Mine also lasted just over a year. These boards run way too hot without a CPU fan. Even though I got a replacement, I have a feeling it'll fail again. Looking into a better cooling kit for it. Will post here if I find a solution
 

Nick2253

Neophyte Sage
Joined
Apr 21, 2014
Messages
1,455
I found the following thread on Reddit:
https://www.reddit.com/r/homelab/comments/4ufa8i/c2750d4i_wont_power_up/
This is very concerning. Can anyone comment on this? I'm wondering if this is new behavior in Freenas 9.10? My problem and and the one mentioned in this thread is identical, even the 15 day timing.
I'm about to get a replacement asrock board, but don't want to use this new board if I'm going to have the same issue.
I'd be hesitant to believe that this is actually the case. The redditor provides no evidence for their claim, and a (admittedly brief) Google search turns up nothing suspect. Anecdotally, there are a number of people here using that motherboard 24/7 who have not run in to problems.

In general, based on my experience with bad motherboards, I'd tend to believe the "bad power" or "excess heat" explanations carries more weight, unless we start seeing some hard evidence otherwise. I have definitely run in to some motherboards that are much more sensitive to bad power than others, which is why I consider a top-end PSU an absolute requirement for whatever system I'm building.
 

Jason Fayre

Junior Member
Joined
Apr 11, 2015
Messages
18
I'd be hesitant to believe that this is actually the case. The redditor provides no evidence for their claim, and a (admittedly brief) Google search turns up nothing suspect. Anecdotally, there are a number of people here using that motherboard 24/7 who have not run in to problems.

In general, based on my experience with bad motherboards, I'd tend to believe the "bad power" or "excess heat" explanations carries more weight, unless we start seeing some hard evidence otherwise. I have definitely run in to some motherboards that are much more sensitive to bad power than others, which is why I consider a top-end PSU an absolute requirement for whatever system I'm building.
Hello Nick,
I completely agree with you, accept for the bug report I referenced in my earlier message. I saw the identical symptoms on my first replacement board from Asrock.
On this third board, I have turned off watchdogd, hopefully that will help.
The bug report is at:
https://bugs.freenas.org/issues/16190
 

nickt

Member
Joined
Feb 27, 2015
Messages
127
Another disappointed C2750D4I user. 15 months later, I've got a dead board. If you're interested, the gory details are here. In short, my issue appears (symptomatically) to correlate well with the BMC flash RAM being rewritten very frequently (and dying). My server is well ventilated and in a cool environment, board and drive temperatures were always very low. (Drives have never exceeded ~32, but typically sit in the low 20s. CPU / board, I can't remember well, but were in the 50 - 60 range, never more).

I have sympathy for the comment about looking to heat / power supply issues first - my intuition would suggest the same thing. But my symptoms seem to support the watchdog / BMC / flash RAM theory well. At times, I am able to get the board going again by changing a configuration setting in BMC, which - presumably - sometimes gets a "good write", and the board can POST again. I've even been able to run a memtest86 solidly for a few hours and run FreeNAS for a few hours. So this rules out many things, in my mind. But with absolute certainty, at some point, the whole things packs it in again, and it refuses to POST.

Interestingly, when the board first failed, the only hint of anything gone wrong in the BMC logs were CPU temperature alarms (all the way up to critical). I strongly doubt that they were legitimate, as it was cool and there was no CPU / IO load at the time. I think instead it was evidence of the BMC losing its mind. The BMC often reports sensor errors in its logs.

The even bigger frustration for me is - being an Australian - the completely non-existent support from ASRockRack global. The US support seems good, but global (out of Taiwan) are doing almost nothing. 3 weeks later, I still haven't been able to start the RMA process.

All rather disappointing.
 

Stux

Wizened Sage
Joined
Jun 2, 2016
Messages
4,222
I think the Asrockrack Avoton boards need to come off the hardware recommendation list.
 

Ericloewe

Not-very-passive-but-aggressive
Moderator
Joined
Feb 15, 2014
Messages
17,431
I think the Asrockrack Avoton boards need to come off the hardware recommendation list.
Yeah, it's what's been holding up my new version of the guide. I was about to finish it when I heard about this... aaaaaand ASRock Rack's response has not been the best, to be honest.
 

Stux

Wizened Sage
Joined
Jun 2, 2016
Messages
4,222
If it were Apple there'd be a class action brewing
 

DaveY

Member
Joined
Dec 1, 2014
Messages
141
...
Interestingly, when the board first failed, the only hint of anything gone wrong in the BMC logs were CPU temperature alarms (all the way up to critical). I strongly doubt that they were legitimate, as it was cool and there was no CPU / IO load at the time. I think instead it was evidence of the BMC losing its mind. The BMC often reports sensor errors in its logs.

I don't doubt the watchdog bug caused a lot of your headaches, but I can tell you from personal experience that your CPU probably overheated and fried some stuff around it. I thought the same of my board when it started having problems. IPMI kept reporting my CPU/motherboard were in the 60's with the system idle. And it will always alert with critical temps when I run Plex. I opened the case and felt the heatsink while it was running and it was cold so I thought the sensors were wrong. After the board died, I decided to remove the heatsink just to see if maybe there was an issue with the CPU. It turns out the thermal pad was all shredded where the cpu and heatsink are suppose to make contact so it left a slight gap. There's about a .1 mm clearance between the cpu and the heatsink and the thermal pad was meant to fill that gap. Without it, your CPU has no way to transfer the heat it generates and it WILL leave your heatsink cold and the CPU red hot.

After I got my replacement board, the first thing I did was remove the heatsink, slapped some arctic silver on there and a copper shim and made sure heatsink/cpu had good contact. I also added a heatsink fan like I mentioned in my other post. The CPU now runs in the 20s when idle and never goes above low 30s. The motherboard still reports 40-55c, but that's understandable as I don't have a fan blowing across it.

Here's my IPMI readings as I type this:
Code:
MB Temperature  | 44.000  | degrees C  | ok  | na  | na  | na  | na  | na  | 80.000 
CPU Temperature  | 28.000  | degrees C  | ok  | na  | na  | na  | 80.000  | 85.000  | 90.000


Anyone with the Avoton board and seeing CPU temps above 40s should double check the cpu/heatsink contact. I really think ASRock needs to fix this silliness. They have a solid product, but now have boards coming back for RMA by the truckloads all because they tried to save a few penny on the thermal compound.
 
Last edited:

DaveY

Member
Joined
Dec 1, 2014
Messages
141
I think the Asrockrack Avoton boards need to come off the hardware recommendation list.

Might be too late. Think of all the FreeNAS Mini users?? I have a feeling we'll be seeing their posts shortly.
 
Status
Not open for further replies.
Top