Is my hard drive on the way out?

nigelc · 12-06-2008, 09:28 PM

You have to make it really, really cold for the resistance to be measurable like > -50c. When I used to fix power supplies and hard drives most faults were cold i.e when it was first powered up. Electronics work better when they are at a constant temperature. The plugs & connectors not expanding and contracting. Dry joints always worse when cold. If the drive had heated up, then the electronics have been over-stressed and will probably fail.

deepsix · 12-07-2008, 01:12 AM

Quote:

Originally Posted by eggbert74

Every once in a while my hard drive will kind of freeze up. Sometimes this will occur when booting up, and other times it'll occur at seemingly random times. When it happens the entire machine locks up, and it'll either come back to life after about 20 seconds, or I'll be forced to do a hard reboot. While the machine is frozen, the hard drive makes a kind of seek noise. Sort of like chhh chhh chhh chhh.

Here's what my kernel log says after this happens.

Code:

Dec  1 11:22:28 pandora kernel: [  735.000049] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
Dec  1 11:22:28 pandora kernel: [  735.000061] ata1.00: cmd 35/00:08:8d:94:c9/00:00:14:00:00/e0 tag 0 dma 4096 out
Dec  1 11:22:28 pandora kernel: [  735.000063]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
Dec  1 11:22:28 pandora kernel: [  735.000067] ata1.00: status: { DRDY }
Dec  1 11:22:28 pandora kernel: [  735.000075] ata1: hard resetting link
Dec  1 11:22:33 pandora kernel: [  740.512011] ata1: link is slow to respond, please be patient (ready=0)
Dec  1 11:22:38 pandora kernel: [  745.048015] ata1: SRST failed (errno=-16)
Dec  1 11:22:38 pandora kernel: [  745.058743] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Dec  1 11:22:38 pandora kernel: [  745.058753] ata1: link online but device misclassified, retrying
Dec  1 11:22:38 pandora kernel: [  745.058759] ata1: hard resetting link
Dec  1 11:22:43 pandora kernel: [  750.572018] ata1: link is slow to respond, please be patient (ready=0)
Dec  1 11:22:48 pandora kernel: [  755.108014] ata1: SRST failed (errno=-16)
Dec  1 11:22:48 pandora kernel: [  755.118767] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Dec  1 11:22:48 pandora kernel: [  755.118783] ata1: link online but device misclassified, retrying
Dec  1 11:22:48 pandora kernel: [  755.118790] ata1: hard resetting link
Dec  1 11:22:53 pandora kernel: [  760.632020] ata1: link is slow to respond, please be patient (ready=0)
Dec  1 11:23:23 pandora kernel: [  790.152010] ata1: SRST failed (errno=-16)
Dec  1 11:23:23 pandora kernel: [  790.162741] ata1: SATA link down (SStatus 21 SControl 300)
Dec  1 11:23:28 pandora kernel: [  795.160026] ata1: hard resetting link
Dec  1 11:23:29 pandora kernel: [  796.532061] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Dec  1 11:23:29 pandora kernel: [  796.565352] ata1.00: configured for UDMA/133
Dec  1 11:23:29 pandora kernel: [  796.565365] ata1: EH complete
Dec  1 11:23:29 pandora kernel: [  796.565868] sd 0:0:0:0: [sda] 976773168 512-byte hardware sectors (500108 MB)
Dec  1 11:23:29 pandora kernel: [  796.566088] sd 0:0:0:0: [sda] Write Protect is off
Dec  1 11:23:29 pandora kernel: [  796.566098] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
Dec  1 11:23:29 pandora kernel: [  796.566482] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

I would suspect that the hard drive is failing, however I dual boot and have no trouble under Vista. Also, for what it's worth, I've run countless HD diagnostic tools and everything checks out ok.

Can anyone shed some light on this? I'd love to know what the log is saying and/or how I can reliably determine if my HD is going to bite the dust.

My specs are:
Dell Inspiron 530 desktop. Intel Q6600 Core 2 Quad w 3GB of RAM. Western Digital 500 GB SATA drive. Running Ubuntu 8.10.

have you looked into the bios?...
some bioses have settings for non dos systems for disk access
I had the same problem at boot on an older machine and I just changed the setting in the bios for other or non dos machines
...also I kept getting incorrect drive allocation with the dos or windows setting.

eggbert74 · 12-08-2008, 07:38 PM

Quote:

This is just my personal experience, but I hate WD HDD's. I have seen several of them fry. I won't use them any more. One other thing. If the computer was cold, then wouldn't the power wires conduct better? The resistance should increase as the wire heats up. I would get another HDD and move my system's over if it were me. However, that is more experience based than anything, so it may mean nothing in your case,

You may be right, but I just have a hard time believing it's the drive. Seems like if the drive was failing the SMART data would give some such indication, or I'd find a run of bad sectors, or at least SOMETHING to tell me the drive was going down.

Quote:

have you looked into the bios?...
some bioses have settings for non dos systems for disk access

I checked, but sadly Dell saw fit to provide no other options in the BIOS other than the bare essentials.

daw617 · 01-14-2009, 07:50 PM

Did you ever figure this out? I'm having a similar problem, on my Dell XPS 420. I had the same exact hard drive and the same problem. However I tried a different hard drive and the problem persisted, so in my case I suspect it may not be a problem with the hard drive itself but something else (SATA controller? I don't know). It also seems to vary depending upon what version of the Linux kernel I use: with a Linux kernel downloaded from kernel.org, configured to minimal settings, and compiled by hand, I'm not experiencing the same quantity of SATA errors.

Details:

https://bugzilla.redhat.com/show_bug.cgi?id=470623

eggbert74 · 01-16-2009, 07:06 PM

daw617,

I was never able to solve it. I finally ended up contacting Dell and they sent me a replacement drive. So far the problem has not occurred with the new drive. I've been running with it for about a month. *Knock on wood*

The replacement drive is the same model WD5000AAKS. However, for what it's worth, it has a different revision number than the old drive. This one ends in B0, the old clicky one was (I think) A0. A problem with the A0 revision maybe? In my case, I'm pretty sure is was a mechanical problem with the drive. It would even occur under windows occasionally.

daw617 · 01-16-2009, 07:15 PM

Cool, glad to hear that it worked out! I guess we probably have a different set of problems.

Incidentally, I recently tried forcing my drive and controller to downgrade to a 1.5Gbps SATA link, rather than the default of 3Gbps. I'm having good luck with this so far -- no crashes or problems so far. I thought I'd write this down in case anyone else has a similar issue and stumbles across this via Google: try forcing the SATA link to 1.5Gbps, either via jumpering the hard drive or by "options libata force=1.5Gbps" in modprobe.d.

edroid · 02-03-2009, 01:49 PM

I encountered this problem with two SATA drives, a Western Digital (model WDC WD1600JS-COM) and a Seagate Barracuda (model ST3160811AS). Both drives worked fine with OpenSUSE 10.2 and Kubuntu 8.04. But when installing openSUSE 11.1, the following reports were repeated several times to the splash-screen while the install kernel booted:

ata1: hard resetting link
ata1: link is slow to respond, please be patient (ready=0)
ata1: SRST failed (errno=-16)
ata1: SATA link up 3.0 Gbps (SStatus 113 SControl 300)
ata1: link online but device misclassified, retrying

The install proceeded, but could only see my 80GB PATA drive. The same thing happened when trying to install Fedora 9 on this system.

Google led me to this thread and setting the drive jumper to 1.5Gbps fixed the problem for both drives.

My questions are:

1. Would all SATA drives have this problem with openSUSE 11.1 on this system?

2. Which is the best solution: using the drive jumper or setting the option?

3. Where/how do I set the option to "force=1.5Gbps"?

daw617 · 02-03-2009, 02:23 PM

Quote:

Originally Posted by edroid

2. Which is the best solution: using the drive jumper or setting the option?

I found the drive jumper easier (if you have a spare jumper laying around): you set it once and never have to think about it again. But either way should work.

Quote:

3. Where/how do I set the option to "force=1.5Gbps"?

Unfortunately it depends upon whether you have libata built into your kernel, or loaded as a module; and if it is loaded as a module, whether it is loaded by your initrd.

If it is built into the kernel: add "libata.force=1.5Gbps" to the list of kernel options on the boot command line (you can do this with Grub when the system starts up and you get to the bootloader). In my /etc/grub.conf I have a line like this:
kernel /vmlinuz-(blah) libata.force=1.5Gbps ro root=UUID=(blah) quiet
(The "(blah)" is omitted stuff.) Be careful to avoid typos.

If it is not built into your kernel: first, add the following line to your /etc/modprobe.conf file (create that file if it does not exist):
options libata force=1.5Gbps
(Warning: You may have a /etc/modprobe.d directory, with a bunch of files in it. You can also create a file in that directory containing just the above line, but due to a bug in initrd, that may not be enough; you will probably also want to add it to /etc/modprobe.conf as well.) Then, re-build your initrd. For instructions, see here: http://www.linuxquestions.org/questi....5gbps-697997/

You can tell whether it is working by rebooting into the new kernel and running "dmesg". If you see a line like this:
ata1: FORCE: PHY spd limit set to 1.5Gbps
then you successfully set the libata option to "force=1.5Gbps".

arnuld · 02-04-2009, 12:20 AM

I had similar kind of problem with my HDDs when I was using XFS filesystem. I reverted back to ext3 and HDDd busy(ness) disappeared. Since you did not have any problems on Windows Vista, what filesystems on Linux you are using ?

NOTE: I think its better to find out the reason before new one goes down too.