3TB hard disk issues on FreeBSD 8.2 RELEASE (freezes)

Z

zphantom

Hello all, quick question I hope you can help with.

I have been using FreeBSD for quite a while (4.7 days), and have always had success with it. I am currently running FreeBSD 8.2 RELEASE on a older computer with a few Si3124 controllers (one PCI, one PCI-X) and 5 hard drives (120GB for system, 1.5TB data drive, and 3 new WD30EZRX 3.0TB data drives). The computer has a 460W power supply, and the 3 3.0TB drives are the "green" drives, which don't take much power.

I have formatted each of the 3.0TB drives using the instructions in: http://forums.freebsd.org/showthread.php?p=76147, with the exception of replacing the gpt command with gpart. I should also note they are jumpered to 3.0Gb speed setting (given the controller I am using)

I have been testing the drives by making a 6.0GB file: dd if=/dev/zero of=file1 bs=1m count=6144 and then copying that file (in a for loop) numerous times.

Sooner or later, the system will freeze (hard lock). I originally thought it was a heat issue, but I have installed a nice 120mm fan and according to smartctl none of the drives cracks 40C.

Does anyone have any thoughts or suggestions. It is random enough of a freeze to point to hardware or driver in my mind. Have I done something incorrect with regards to the formatting?

Note: These messages do show up in my /var/log/messages:
Code: 
GEOM: ad6: the secondary GPT table is corrupt or invalid.
for all three drives.

Thanks in advance for the help.

Mike.

Code: 
atapci0: <SiI 3124 SATA300 controller> port 0xdcf0-0xdcff mem 0xff6ffc00-0xff6ffc7f,0xff6f0000-0xff6f7fff irq 25 at device 12.0 on pci3
atapci0: [ITHREAD]
ata2: <ATA channel 0> on atapci0
ata2: [ITHREAD]
ata3: <ATA channel 1> on atapci0
ata3: [ITHREAD]
ata4: <ATA channel 2> on atapci0
ata4: [ITHREAD]
ata5: <ATA channel 3> on atapci0
ata5: [ITHREAD]
atapci1: <SiI 3124 SATA300 controller> port 0xccf0-0xccff mem 0xff2ff400-0xff2ff47f,0xff2f0000-0xff2f7fff irq 22 at device 14.0 on pci5
atapci1: [ITHREAD]
ata6: <ATA channel 0> on atapci1
ata6: [ITHREAD]
ata7: <ATA channel 1> on atapci1
ata7: [ITHREAD]
ata8: <ATA channel 2> on atapci1
ata8: [ITHREAD]
ata9: <ATA channel 3> on atapci1
ata9: [ITHREAD]
atapci2: <Intel ICH4 UDMA100 controller> port 0x1f0-0x1f7,0x3f6,0x170-0x177,0x376,0xffa0-0xffaf at device 31.1 on pci0
ata0: <ATA channel 0> on atapci2
ata0: [ITHREAD]
ata1: <ATA channel 1> on atapci2
ata1: [ITHREAD]
ad0: 117800MB <IC35L120AVVA07 0 VA6OA52A> at ata0-master UDMA100 
acd0: DVDR <SONY DVD RW DRU-500A/2.0c> at ata1-master UDMA33 
ad6: 2861588MB <WDC WD30EZRX-00MMMB0 80.00A80> at ata3-master UDMA100 SATA 3Gb/s
ad8: 2861588MB <WDC WD30EZRX-00MMMB0 80.00A80> at ata4-master UDMA100 SATA 3Gb/s
ad10: 2861588MB <WDC WD30EZRX-00MMMB0 80.00A80> at ata5-master UDMA100 SATA 3Gb/s
ad18: 1430799MB <Seagate ST31500341AS CC1H> at ata9-master UDMA100 SATA 3Gb/s
 
You may try to use new siis(4) driver for these SiI3124 controllers. It should be at least faster, but may be it also give some more info, especially if booted with verbose kernel messages.
 
Since it's green, you may also want to check Load Cycle Count in smartctl. I have green on the desktop and the drive used to park it's head while i'm working on that PC.
 
Solved

The main issue with my setup was "long" SATA cables. I had two of the three
hard drives on longer (1m) SATA cables.

Once I switched to the siis driver, I would occasionally get
timeout messages from the kernel on those devices. I was able to replace
the longer cables with shorter standard cables, and was able to torture
the system to where I believe everything is working properly.

Since it's green, you may also want to check Load Cycle Count in smartctl.
Click to expand...

hedgehog, this is a good suggestion, luckily for me these are data drives, so they
are only parked when I am not accessing them, and they don't get accessed often.
My LCC is small.

Thanks again.
 
Doesn't the WD Caviar Green drives go to sleep after 8 seconds every time of idling which would explain the freezes?
 
What kind of CPU/GPU are you using in this computer? Since you say the system was hard locking, the first thing that came to mind was that with 5 HD's perhaps the computer just isn't getting enough power from the PSU.

With regards to the head parking on the Green drives, I use a 2TB version of the WD Green drives (WD20EARS) and I had a big problem with what seemed like freezing/massive system slowdown in FreeBSD.

I used the WDidle3 program (http://support.wdc.com/product/download.asp?groupid=609&sid=113 and changed the idle time of the drive from 8 seconds (default) to 5 minutes (IIRC) and the problem has disappeared.

The page I linked to mentions that the program is only for certain drives, but there was a thread on the WD Community Forums (that I can't seem to find at the moment), where it was confirmed by someone working for WD that the program was fine to use on other WD Green drives.
 
The WD drives do spin down, however since they are data drives only (I have the main slices on an older IDE drive), it does not affect anything. They are in use when I access the data, and otherwise spun down.

As for the hard locking, it was traced to a bad stick of RAM. Soon as I replaced it, I have been up for a month with no issues.

By the way, not that I think it matters but the GPU is a ATI Radeon (9700 I think)

Thanks again.
 
You must log in or register to reply here.
Back
Top