TRAP 28 - FreeNAS stops working.

[joeschmuck]

Running Build 8r6579M (added ataidle change to add -I parameter), 4 drives in RAIDZ2, AIO = ON.

While transferring continuous data from my main windows computer to the FreeNAS drive I receive a TRAP Number 28, Current process = panic: machine check trap, 7501 (smbd), cpuid=1. The previous time this occurred I was running all 4 drives as RAIDZ.

I also got this a second time with a 2219 (smbd) value. I tried to find info about the trap 28 error with respect to FreeBSD but failed.

FreeNAS Computer setup:
ASRock G41M-LE, E8500 CPU, 4GB RAM, 4 Samsung F4 drives, 2GB USB thumb drive, and a CD-RW drive to load the iso.

I have run Memtest 2.10 on my RAM and it never failed (3 complete passes).

I will try again with the official Beta release but I doubt that will change the outcome.

Any suggestions?

-Mark

 

[ixdwhite]

Machine check panics usually indicate something's wrong with your CPU or motherboard. There should be a few lines above the panic starting with 'MCA' that should help identify the faulty component. If your CPU is still under warranty you might want to just swap that now. Also make sure you disable all overclocking and reset BIOS settings to defaults.

Also try memtest86+ (http://www.memtest.org/) for something written for modern CPUs. memtest 2.10 has been unmaintained for _years_.

 

[joeschmuck]

Thanks, I'll give that a try. As for the error message, I'll take a digital photo the next several times they occur to try and piece together the problem. As for the CPU, this is my old E8500. It works like a champ, or so I think it does. The MB is new. Maybe I'll run Prime95 on it to give it a good testing as well I'll run the latest MemTest. There is no overclocking going on, and to be honest, I'm thinking about underclocking the CPU to reduce energy consumption. I have had it down quite low before but since running into this problem I have restored it to the standard defaults. Maybe the MB needs a little tweaking.

I suppose the info after "MCA" is obvious what the problem component is? I know if can't be that simple.

-Mark

 

[joeschmuck]

Error Message Photo

Here is the error message and I do not see MCA anywhere... Any advice?



I did turn off the power on the FreeNAS box and removed the USB key in case there is some valuable data on it to assist in isolating the problem (not sure if there is an error log). For this error I was using build version 8r6658 which might I say seems to perform very well except for my issue. I have not run the RAM test using the newer MemTest but will and I need to find a bootable CPU stress tester.

My RAIDZ was all four drives. I know that before I went to bed, for about 90 minutes I was transferring data at full bandwidth of my Gbit network and it was really looking good but then when I go up this morning, failure.

 

[joeschmuck]

I have been testing for about the past 7 hours, 3 of it running MEMTest and stressing the CPU, neither of which twitched at the sign of tuff testing (all passed without incident). I decided to replace the flash drive with a 4 year old one (Adata PD7) which is a fast drive (20MB/sec read, 10MB/sec write on random 512K files, or 22MB/sec read & 14MB/sec write sequential). I have been flooding the data through e FreeNAS box and so far no issues. I will let it run all night and see what it looks like in the morning but it's looking up. If it stays working I would not recommend PNY cheap drives although it would have only failed while using it in the FreeNAS box.