Hello,

 

Hopefully someone here will know what's causing my problem and how to fix it.

 

The hardware is:

Motherboard: ASUS P5B-VM DO

CPU: Pentium D 805

Sata Controller (2): PROMISE SATA300 TX4 PCI SATA II Controller

Hard Drives (5 for now): SAMSUNG SpinPoint T Series HD501LJ 500GB

PSU: PC Power & Cooling Silencer 610 EPS12V EPS12V 610W Continuous

Hot swap bays (3): Athena Power BP-SATA3051B 3 x 5.25" Bays to 5 x 3.5" HD (SATA) Back Plane

Ram: CORSAIR ValueSelect 1GB (2 x 512MB) 240-Pin DDR2 SDRAM DDR2 533 (PC2 4200) Dual Channel Kit

Flash Drive: CORSAIR 512MB Flash Drive (USB2.0 Portable) Model CMFUSB2.0-512

 

After several hours to a couple of days my parity drive will start finding errors and clicking.  It tends to happen mostly, but not always, when transferring files to disk 2, regardless of which disk/slot is disk 2.  It will steadily increase in errors on the main page for the parity drive until it quits responding.  I've tried different drives for parity, different slots in the backplane and even different flash drives.  Nothing seems to change this.  I've tried unRAID 4.1, 4.2 Beta 4 and now 4.2.1.  4.2.1 is the only version that when this happened it did not stop responding and I was able to shut it down.  It took it almost 30 minutes for it to stop the array though.  I did grab the log before trying to stop the array on the off chance it was going to quit responding.  It weighs in at 1.19MB and can be found here.  I've tested all of the hardware under Windows and I know it is good.

 

4.2.1 is the first version I didn't have to wipe all of the disks and start over with.  The server came back up today when I powered it up and all looks well.  I'm currently moving 400+GB over to disk 2 to see if it repeats itself.  Any hints or suggestions are welcome.

 

I also have a question.  How often should one run a parity check?  It currently shows my last parity check over two days ago and says it is good.

 

Thanks.

I had the same problem with the very same hard drive as you (Samsung HD501LJ).

 

Had 2 BRAND NEW drives (as above) in my array and heard a clicking sound.  In the end I got so fed up with it I took them both out and did a FULL (not QUICK) format under windows (one at a time) and then did a FULL check for errors.

 

When I woke up this morning one of the drives had errors reported so its now going back to ebuyer.

 

I would suggest that if you can, you should try a format and error check on your drives.

 

Your syslog looks fine until 4:30am.  From the syslog, it appears you booted at 8:52pm, assigned Disk2 and started it clearing, which finished at 11:03pm.  Around 4:09am,  you formatted Disk2 which finished at 4:10am.  Then at 4:31am, the errors start and continue to the end of the syslog.  There are quiet periods, but the error clusters are between 4:31am and 6:23am, between 4:54pm and 5:14pm, between 9:30pm and 11:13pm, and from 4:16am to the end of the syslog at 4:20am.  I mention these times in case you can associate the times with whatever actions or copies you may have been making.

 

The errors are all concerning the parity drive, a SAMSUNG_HD501LJ with serial # ending ...100.  The #1 suspect here would be that this is a bad drive, especially since you mentioned clicking.  There is probably no more feared warning sign of a drive going bad than if it begins clicking.  If the drive had data, it would need to be copied off immediately, even before further diagnostic tests are run on that drive.  A parity drive contains no data though, but you will temporarily lose the protection of the data on the rest of your unRAID data drives.  A clicking drive is probably not repairable.

 

Other possible suspects are bad power, bad cable, or bad port.  You can try connecting to a different port, and you can replace that cable.  Your PSU is about as good as it gets, but even the best could go bad.

 

It might be useful to examine the SMART data from this drive.  This thread http://lime-technology.com/forum/index.php?topic=664.0 explains how to obtain SMART info from the drive.  The download is missing, so I've attached smartctl.zip, containing smartctl with the flags set correctly.  Unpack it to the flash drive, then at a Telnet console, type:  /boot/smartctl -a -d ata /dev/sda.  Change sda as appropriate.

 

Well, attachments do not seem to be working currently, could not attach smartctl.zip, size is 62k.

 

I'm not worried about the data right now as all of it can be replaced.  I've been testing out the basic version with the intention of purchasing the pro version once I was sure of my understanding and ability to run it (and remember all that lovely linux stuff I forgot about 10 years ago).  Somewhere in there I had started copying 400+GB to disk 2 before going to bed.  The day before (I have that log also, but no errors occurred) I had copied roughly the same amount to disk 1.

 

I've tested all of the equipment under Windows and everything works fine.  I've filled the drives and wiped them clean, no errors.  The funny thing is, it doesn't matter which drive is the parity drive nor which slot it is in on the backplane, the parity drive is always the one this happens to.  It starts generating errors sometime during the copying to disk 2 phase (regardless of which disk is disk 2 and which slot it is in).

 

4.2.1 did do one thing for me though that none of the other versions would.  Previously I would have to wipe the drives in Windows or unRAID would not finish the boot phase.  I've currently got it rebuilding parity with the data currently on it to a WD 500GB drive.  I'm wondering if something in the hardware/software mix just doesn't like the Samsung drives.  Samsung has an HD utility, so I guess I'll hook up a monitor and keyboard to that system and see what it tells me.  I'll also keep an eye out for that download to reappear also.

 

Thanks for the help.

Thought I'd do a follow up post and let everyone know what I've found (and haven't found).  No matter which of my 15 bays I place the Samsung drives, the parity drive errors out when writing to disk 2.  I've since placed a Western Digital drive in as the parity and all is well.  I've filled disk 2 with no errors being generated.  I think the problem lies in some strange hardware compatibility problem between the Samsung drives and bios settings or back planes.

 

All works well now, I've simulated breaking a data drive and the parity drive (separately) and recovered with no problems.  I just finished ordering the pro version.