Before I rip my new ( but failing) system apart...

[NeilP]

Before I rip my new ( but failing) system apart can some one tell me what the issue may be.

I think I already know but maybe if someone with a bit more knowledge looked they could tell me. I will not even say what I think it is, just give the symptoms and the setup

Gigabit M/b with 3ghz processor..not sure of exact spec
2Gb RAM
365 Watt PSU

4 x SATA HDD's of various types (2x500gb, 1x1TB, 1x750gb)
1 x IDE 750 Gb
1x CDROM

1x CPU fan
1x Front case fan
1x rear case fan

It seems to run for a while , then cuts out, last failure was grabbed from the screen with the iPhone..it ran for 11 minutes

 

[ProtoSD]

First guess is memory, but you also didn't mention how your drives were configured, UFS / ZFS.

 

[NeilP]

Ah
ZFS, all individual drives.
2g RAM on the only two slots

Why do you say memory? in what respect?.

I am still keeping my thoughts back, so as not to lead people down that path. Am after resh idea liek your memory idea.

It has been running on/off for a few days and seem to randomly stop

 

[NeilP]

MB is
http://www.msi.com/product/mb/945GM3-F.html#/?div=Basic

 

[NeilP]

Solved ???

Think it is solved.
My thoughts were:
PSU was underpowered for the number of HDD's, and the three fans.
Possible overheating in small case.

I thought I would re build to a bigger case and add a second PSU, to run the HDD's and fans.

BUT look what I found on dismantling, I guess this is playing a part in this.

 

[NeilP]

No not solved

Well another day wasted...I started at about 0930, and it is now 2100 and I have not stopped working on this bloody computer.
Stripped out the mainboard and fitted a new 4 pin fitting which I removed from another board I had.

Rebuilt the machine in a new bigger case, scaveneging parts from at lest 3 other boxes to get it all to fit.

Added a second PSU, case/hdd/cpu temp monitor and fan speed controller for the rear case fan.
Also fitted two fans near the HDD stack to get extra airflow over them.

Got it running about 2 hours ago, but now good still crashing regularly after 7 to 10 minutes, sometime the same eror screen on the console as posted earlier sometimes as in the next picture

 

[ben]

What FreeNAS Version? There were known kernel panics related to plugin jails in the 8.2 betas. I personally found that rebooting while leaving the plugin service enabled led to lots of panics, but I believe all of those are taken care of in RC1 (out today).

 

[NeilP]

Oh sorry, bizarre, i am sure i posted system spec. but on checking you are correct, totally forgot that...umm sorry ..a basic error I know, should always post full spec.
8.04 release p3 just downloaded about 2 or three days ago


FreeNAS Build FreeNAS-8.0.4-RELEASE-p3-x86 (11703)
Platform Intel(R) Pentium(R) D CPU 3.00GHz
Memory 2027MB
Mac 10.6.8

 

[paleoN]

I'd be looking your motherboard at this point. Though did you ever test out your RAM?

 

[NeilP]

Ah well I guess I wont even check it..I'll just scrap it and look for another .

Cant really test the RAM, do not have another machine that it will fit on ,and I am not going to start installing windows or ubuntu or similar on it just to test the RAM

It was a new machine about 2 years ago, used for about 6 months then the owners went off to Australia, took jsut the HDD's with them, it has been sat here ever since.

yep time to scrap it



Weird that it works for a bit then stops. but I have wasted enough time. Just had two week holiday, and have spent the last 4 days solid just working on this bloody thing.
Time to give up and wait for another machine to come along

 

[paleoN]

Cant really test the RAM, do not have another machine that it will fit on ,and I am not going to start installing windows or ubuntu or similar on it just to test the RAM

No installation necessary.

I'd still run memtest86+ on it overnight. Keep in mind it tests your entire memory subsystem and not just the RAM in isolation.

Burn it to a CD or USB drive.

 

[NeilP]

Hey great, thanks will give that a go tomorrow, am in bed now

Cheers for the help

 

[NeilP]

That jsut sows what stopping and having a break can do, and how easy it is to get fixated on a job.
I have a memtest CD any way. found it in my 'CD Tools' wallet..an old version so have grabbed the new one, thanks for the link.
My old FreeNAS 0.7 box was so reliable I have not done any work like this for two years or more,

Test running now
.

It has identified the memory as
DDR2-333

that is odd because when the Machine boots, the BIOS screen shows it has 667mhz
According to the mainboard manual, this board only accepts 400, 533 or 667 RAM..

The online spec for the memory shows it as 667 too.

memtest shows the FSB as running at 199Mhz.

chipset is intel i945P/G ECC is disabled
CPU Pentium D 299sMhz
L1 cache 16k
l@ cache2048K
Nil Level 3 cache

Will leave test running all day and see what transpires, but any comments on the spec I posted regarding compatibility with FreeNAS 8 appreciated

 

[NeilP]

Well "memtest" completed, and all OK on the memory front

Just tried running with just one IDE hard drive connected and no SATA, and it still crashes.

Will create new volumes on the drives in UFS rather than ZFS and see if that changes things

otherwise any other suggestions?

 

[NeilP]

I think AFP is the problem

Nothing conclusive yet..but I think I have found the culprit.

Still crashes with UFS volumes...but

sharing with CIFS rather than AFP and it seems far more stable.
Chronosync has just passed the 2gb copied to the IDE drive with over 12 minutes run time with not a single error message appearing on teh directly plugged in console screen.

I am guessing it is an AFP issue, so will just go back to use sharing via CIFS, AFP seems broken.

Am going out the house now, will leave it run and see if it is still going when i get back

 

[NeilP]

Yep it must have been AFP..been now for 3hrs 45 and copied over 30 gig of data to it, and still OK

 

[NeilP]

Yep it is AFP that crashes it.

After over 5 hrs running and 60gig or more sync'd to it, still running strong

Enable AFP and within seconds of Mac Finder trying to log in it crashes.

Going to stick with CIFS, AFP seems broken.

 

[paleoN]

Yep it is AFP that crashes it.

Is it still crashing with Fatal Trap 12? If so you have a hardware issue. As in malfunctioning hardware or unsupported hardware that doesn't play nice with FreeBSD. I wouldn't count on CIFs keeping it from crashing again. It is a bit strange that AFP exposes it so readily.

What kind of NIC do you have?

 

[NeilP]

what ever is built in to the motherboard.
I could see your point about it being hardware if I had not had it running so well with CIFS for this many hours, where AFP crashes it within minutes.
Will leave it up over night doing my iTunes library backup of about 200gig and see how it holds out..best I had before with AFP was about 12 minutes.

from the manufacturers page http://www.msi.com/product/mb/945GM3-F.html#?div=Detail


LAN

• Realtek® RTL8110SC / 8100C
- Supports 10 / 100 / 1000 Mb/s (1000Mb/s on 8100SC only).
- Compliane with PCI 2.2.
- Supports ACPI Power Management.

 

[Stephens]

It has identified the memory as DDR2-333. that is odd because when the Machine boots, the BIOS screen shows it has 667mhz
According to the mainboard manual, this board only accepts 400, 533 or 667 RAM.. The online spec for the memory shows it as 667 too.

Not as odd as it might seem at first. The "DDR2" portion of the memory spec means that 2x the 333mhz (rounded) number you see is 667mhz (rounded). There are multiple ways of referring to memory, such as PC2-5300. Your specs are OK.