August 29, 200817 yr I'm getting disk errors on a frequent basis after installing a new drive. This often leaves the array unstable and I'm unable to write to it. I'm on 4.3.3 - using 4 Samsung 750GB HD753LJ and two Western Digital 10EACS (Green) Drives. These are all SATA drives. When the error happened I tried to run a smartctl on the WD drives. Here's what happened. ------------------------------- root@DIRECTO_MEDIA:~# smartctl -a -d ata /dev/sde smartctl version 5.36 [i486-slackware-linux-gnu] Copyright © 2002-6 Bruce Allen Home page is http://smartmontools.sourceforge.net/ Smartctl: Device Read Identity Failed (not an ATA/ATAPI device) A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options. -------------------------------- After these errors happen, I have to do a hard reset (I can't shutdown from the web GUI or from the command line). After the reset, I was able to run SMART commands on the drives and it passed. A parity sync will happen and everything will be fine, until this happens again (it usually rears its head when I do a write). Can anyone smarter than me help me out here? What's going on? Sometimes the array will work fine for a couple of days, and then just kaput! Attached is my syslog. Thanks in advance for anyone who might be able to help out!
August 29, 200817 yr Your system booted fine, recovered from a previous crash, and finished a parity check without any issues at all, at a very good average speed of 64.2MB/s! But it appears to have lost communications with the WD's at 3:34am, causing hard resets, but unsuccessful in accessing them again. The rest of the errors are just followup to the initial loss of the drives. How are the WD's connected, to what disk controller? Can you give a little detail as to the motherboard and disk controllers? Although there is no direct evidence, this appears to me to be a problem with spin up or spin down. You can verify that by disabling spin down for awhile. There have unfortunately been reports of some drives that refuse to 'behave' with spin down correctly, perhaps incompatible combinations of drive firmware, motherboard, and/or disk controller, and/or driver. I personally have had one drive that, if allowed to spin down, would never spin back up, and result in many errors, and be disabled. On reboot, it would come up fine, but parity or the drive would have to be 'rebuilt'. I recommend going into the BIOS settings and changing to AHCI mode, which will result in a higher performance driver, better suited to simultaneous access. A very minor item: while in the BIOS menus, you might disable the serial and parallel ports, as you probably don't use them on this machine, and it saves unRAID from setting up IRQ's and drivers for them.
August 29, 200817 yr Author Thanks for the insights! It's an Asus P5E VM DO, with I believe an Intel ICH9D0 Sata Controller. From the syslog, can you tell that I'm not running in AHCI mode? My box is setup without a monitor so it's a bit of a pain if I need to change something in the bios... Thanks again for the help, and let me know if the controller info is relevant.
August 29, 200817 yr Thanks for identifying the board. From the syslog, you are not configured for AHCI. It is probably in IDE mode. scsi1 : ata_piix scsi2 : ata_piix ... scsi3 : ata_piix scsi4 : ata_piix It is clear that it is using the standard Intel ata_piix driver. If it were using AHCI, then those lines would be: scsi1 : ahci. It's not vital to change it though, but since we don't know the exact source of the problem, an incompatibility with the driver (among other things) is still a possibility, and changing it to AHCI would change the driver used, and also some of the communication methods with the drives. It's more important to test spin down, by disabling it and running for awhile. Later and more drastically, try clicking the Spin Down button, then testing access to the drives.
Archived
This topic is now archived and is closed to further replies.