Maxtor IDE HD = No good on Linux?
 

Once, for the hell of it, I tried to install an AMD64 version of Linux on an extra HD I have in this computer. I'm running an Opteron CPU so I figured that there's no reason not to use an AMD64 version.

Strange thing is, it didn't matter what version I tried to install (FC4, FC5, Slamd64 10.2b(unoff. 64-bit Slackware port), or Gentoo 2006.0). I got I/O errors whenever I tried to access the only Maxtor HD I have, and fdisk can't read it. However I was able to access the Seagate IDE HD that was on the same IDE channel.

Is this common behavior? Granted, with an experience like this, and many people telling me that Maxtor is no good, I'm only going to buy WD and Seagate drives now, but is every Maxtor IDE HD this way?

Whoever would've thought that something simple like a HD, using the IDE interface (which has been around since time out of mind), might now work with Linux? Linux can even access this WD SATA drive I have, and SATA hasn't been around long at all. But it can't access a Maxtor IDE drive??


Plus suppose I took out the HD, and put it into a USB enclosure. Could Linux access it then?

Of course, the Maxtor HD works perfectly under Windows...


Ben

if you have a regular x86 32 bit PC it wont work, you need a regular x86 distro to match it...

Of course, but I said that I have an Opteron based PC, which is AMD64 (x86-64).

Perhaps I wasn't clear. I can boot from the installation CD and start the installer. However, the startup messages complain about I/O Errors on the HD which is the Maxtor HD. It never happens to the other IDE HD I have.

I've already established that to install Linux, I'd have to buy another HD to but into my system. I was just wondering if anyone else had issues with Maxtor HDs.

I just have to say that, if this is a widespread issue, Maxtor better get their act together.


Ben

I have Linux running on serveral systems with Maxtor drives in them. Are you certain that it is jumpered correctly? Have you downloaded Maxtor diagnostic and checked out the drive with it? Hard Drives do go bad and that has nothing to do with Linux.

have you dis-abled the EZ-Bios feature ....... newer maxtor drives have this enabled by default when they are bought new and can be dis-abled by the user.

It works perfectly fine under Windows, so I can't see how it's a configuration problem. This is precisely what makes no sense. It works fine under Windows, but under Linux I get I/O errors??

I'll see if the Maxtor diagnostic tells me anything (I didn't even know such a thing existed, I bought the drive OEM).


To 320mb:

What's EZBios?


Ben

Although I don't happen to use Maxtor drives, I have in the past, and they work just fine with Linux, so I seriously doubt that's the issue.

Personally, from what you've described, I was also thinking it may be a jumper issue. Despite what happens with Windows, I'd take another look at the jumpers on each hard drive, and while you're at it verify that the power cable connections are properly seated. Along these lines, I would also make sure that each drive was specfically jumpered to be either the master or slave - don't use cable select.

Finally, please post your hard drive layout (eg, "primary master = Seagate 60G")

Primary master = /dev/hda = [brand] [model]
Primary slave = /dev/hdb = [brand] [model]
Secondary master = /dev/hdc = [brand] [model]
Secondary slave = /dev/hdd = [brand] [model]

as well as the output from running
note that's a lowercase "L"

To Nlinecomputers:

Last night I tried the Maxtor diagnostic thing, and strangely enough I did get an error that stated something about the drive and how it's going to fail soon (very odd, since it still works fine on Windows, and Disk Manager doesn't show anything amiss). To be safe I copied the data to a different HD, though it now means I have limited space left (excluding the 40GB Maxtor).

To J.W.:

It's not a jumper issue. I assembled this particular PC myself, and I never use cable select. I was 99% sure I wasn't using that, and indeed I was right (I checked). Plus if you could see it the slave HD is attached to the "Primary" connector, and vice versa.

If I used Linux, the device IDs would correspond to different devices like this:

/dev/hda = Maxtor 6E040L0 40GB U/133 2MB
/dev/hdb = Seagate ST380013A 80GB U/100 8MB
/dev/hdc = NEC 3550A DVD+/-RW U/33 2MB
/dev/hdd = Nothing
/dev/hde = Nothing, and I have no IDE channels in addition to the two on my motherboard.
/dev/sda = WD WD800JD 80GB SATA1 8MB (This is the boot and system drive), attached to an onboard SATA port.
/dev/sdb = Nothing, and no other SATA devices or SATA Host Adapters/RAID controllers.

For the sake of brevity, I left out others, since I don't think they're important to the problem.

Also, my motherboard does have an option for software SATA RAID, but when I built the PC I disabled it via the BIOS.

As far as the output you want, I'll have to boot off of a Linux CD and try to grab the output into a text file, since I can't do it now (I'm currently using XPPro32 SP2).

I can tell you this, however: If I boot Linux off of a CD, the verbose messages near the end complain about I/O Errors on hda, which is the Maxtor drive. And if I run "fdisk /dev/hda" I get a complaint that fdisk can't read the HD. I get no such complaint if I type "fdisk /dev/hdb," which corresponds to the Seagate drive on the same channel (so at least I know the IDE channel is still good).

I'll make the output soon (though, so much for my uptime...which has been amazingly good for an XP box).


Ben

If you disconnect the seagate drive how does it perform then? To me this sounds like a bad drive. Windows can be more tolerant of faulty drives then Linux is. This isn't a bad thing IMO as you'll get more warning to be able to move the drive off then in Windows.

Maxtor has good diagnostics. If it says the drive is bad then I'd say it is. Many Maxtors have 3 yr warranties. I'd try and get it replaced.

Also note that Widnows can only check the areas of the drive formated for Windows. So if your Linux partition has the bad spots Windows can't know that. Hence the need for running the diagnostic.

To J.W.:

I hope this message is useful, since it took me ages to get it (I wrote it down by hand, since I couldn't figure out how to funnel it into a text file on a floppy, and now I'm typing what I wrote).

hda: dma_intr: status=0x51 { DriveReady Seekcomplete Error }
hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=26, sector=0
ide: failed opcode was: unknown
end_request: I/O error, dev hda, sector 0
Buffer I/O error on device hda, logical block 0

I also saw the same thing repeated again for four more times, except with these sector/logical block pairs: 8/1, 16/2, 24/3, and 32/4.

After that I saw data which stated certain characteristics about the other HDs I have (partition scheme, which partitions are bootable, FS type, etc.)

To Nlinecomputers:

Currently, all possible hard disk space is consumed with partitions formatted with NTFS. I got the errors when I tried booting Linux from a CD and running fdisk so I can make a Linux partition. I've never actually installed Linux on this computer (yet), since the HD I wanted to use is apparently acting up.

How would disconnecting the Seagate drive make a difference? Are Maxtor drives infamous like those old Conner drives that only like to work on the same IDE channel as another drive of that brand?

And unfortunately, I bought the drive OEM, which I understand means no warranty from Maxtor to the end user (I think?). And the return policy of where I bought it is long gone, since I bought the drive about a couple years ago.

Also, I just tried running chkdsk on the drive (since there's just one giant partition, and it did claim that there are a few bad sectors, about 44KB worth). I know I really should use that program every now and then, though I won't lie about the fact I haven't done that for months since until recently I was extremely busy.


Ben

I suspect that your problem is NTFS. I haven't installed any other NTFS support than is available from the Debian Distro but I have found NTFS access to be very flakey.

I think that if you work on the partitions in windows and create a free space then you will be able to install Linux in the unused space.

I don't know of any partition resizers for NTFS as I haven't needed to look for one up to now.

Come to think of it, I've never owned a hard disk where a disk checking program reported more than zero bad sectors or blocks. Seeing that this hard disk DOES seem to have some bad sectors, how long can I expect this drive to last? I've already copied any data I deemed to be important, so if it died tomorrow it won't matter. I just wonder what sort of time frame I have here.

To Davecefai:

How is accessing NTFS data an issue? I was trying to run fdisk on a hard drive, so I could change the partition scheme. I was not trying to access the data itself. I was able to run fdisk on the other IDE hard disk (the Seagate) I have, though I didn't make or want to make changes on that, and this other hard disk has one single giant NTFS partition on it.

Anyway I don't trust resizers. If I want to make a change I copy the data somewhere else, make the changes, format clean, and copy the data back.

However, thanks for your input.



Ben

Years of experience has taught me that ALL drives have there quirks even different model drives from the same brand. I've seen electrical problems occur that make it look like your drive is going bad when in reality it is the other drive. When diagnosing problems on drives it's best to isolate the drive as much as possible to eliminate any potential problem. I also try to move the drive to a known good computer on the chance the problem might be the controller on the mobo.

This led me to understand that you already have the drive formatted with NTFS.

There should be an option in your HD utility which will low-level reformat the drive - they don't call it that though. It's a long test and end up with a positive "chuck out" or "all is now OK" message. This has worked well for me.