HDD errors!

[Agent007]

hi all,

I was going throught the logs and came across these errors. What do they mean?
This is a brand new disc, 2 days old. 40GB supports UDMA mode 4.

Following is the info:-

/dev/hda:
 multcount    = 16 (on)
 IO_support   =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 4865/255/63, sectors = 78165360, start = 0


thanks & rgds,
007


 

quote:
Oct 30 22:00:37 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Oct 30 22:00:37 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Oct 30 22:00:37 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Oct 30 22:00:37 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Oct 30 22:07:27 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Oct 30 22:07:27 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Oct 31 22:53:10 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Oct 31 22:53:10 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Oct 31 23:25:38 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Oct 31 23:25:38 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Nov  1 18:50:23 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Nov  1 18:50:23 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Nov  1 19:02:03 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Nov  1 19:02:03 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Nov  1 19:02:03 localhost kernel: hda: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Nov  1 19:02:03 localhost kernel: hda: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Nov  1 19:02:25 localhost login(pam_unix)[1510]: session opened for user root by


LOGIN(uid=0)

[voidmain]

I get these messages on old drives when they are about ready to die. I believe it is caused by bad sectors. If you can figure out exactly where the bad spot is you can partition around it and get longer life out of your drive. I find when you place data over this area with the bad sectors the app will come to a crawl when trying to read data from this sector until it is either successful or times out. Might also try reformatting with a "-c" (check for bad blocks). You can also pass the format command a list of bad blocks so it will format around it.

Since it is a new drive I would think about running some disk diagnostics on it and if it finds problems get it replaced. Nothing good has ever come from a drive I have had when you start seeing these messages.

[pkd_lives]

It looks to be failing a CRC and that is always bad. This error can be caused by corruption of files, it can be the hardrive failing in many ways - including physical deteriation. Generally a disk that fails CRC is trash ready.

Unless you can find out the specific cause - but I warn you trying to keep a CRC damaged disk going can be hell - it WILL make your system unstable.

I would suggest you get the disk replaced, and I would suggest you try going direct to the manufacturer. You will need to find the damaged files - if they exist and then reinstall them, which may well mean a complete reinstallation of the data. Out of interest which OS are you using?

If it is a new drive it was probably tuned slightly bad, maybe not bad enough to be spoted by the manufacturer.

It is possible to save these disks, but CRC generally comes back to haunt. This is low level stuff and is not really attributable in any way to a specific operating system, although I have found OS/2 Warp 4 does do this a lot to floppy disks, possibly as a result of using HPFS. And I have my suspicons about NTFS.

[ November 01, 2002: Message edited by: pkd ]

[KernelPanic]

so IBM or Western Digital  

[Agent007]

Hi guys,

Thanks for the info. Btw, I'm on RedHat 8.0..I changed the drive cable to ide1, it was on ide0 first and the prob seems to have gone, atleast for now. This is a new disc, the previous one of the same make (seagate) crashed after a period of six months. Do u guys think that this disc is going too? How come?

thanks,
007

[voidmain]

There are some drive types, and some batches of drives within the same drive types that are just crap or have manufacturing defects. I don't know if this is one of them or if you have a bad cable. My suspicion is that you have a bad cable if moving it seemed to have an effect on the drive. Also make sure you don't have any master/slave conflicts with other drives.

[pkd_lives]

Okay you switched IDE port and the error vanished. I am pretty sure this is hardware related.

It could be the drive - I am guessing on the bad hardrive because you are getting CRC errors. By the time any hardrive is reaching Cyclic Redundancy Checks, then usually the connection hardware has been proven. However that is not always the case. I don't know the order of tasks from startup - It's probable that only a couple of people in the world actually know this, and that is most likely guess work.

So it's a process of elimination. You changed the port:

Switch back and see if the problem replicates. Use the same cable if possible. Switch back and see if the problem vanishes switching from prot to port. If it continually follows. That can indicate a bad port (which could mean a mobo failure).

If the problem comes and goes on both ports, then suspect the cable. If you have another cable try it out, and see if you can get it to repeat the fault. If the cable switch eliminates the prob, then dump the old cable. It's always good to suspect ribbon cable, it has failed on me many times - people like to crease that stuff like paper, when it should be allowed to curl in a more natural manner.

Now it could be the hardrive, the motherboard or the cable, and it could be intermittent (which means you may be fault finding this for a long time), if it is then I am afraid you will just have to keep watching it until something fails properly.

I doubt this is software related, but it is always possible.

You can always try a reinstall. Start up minimal and add a program, and keep checking and if it switches on at a certain programme's installation then bob's your uncle.

[Agent007]

I forgot to ask, what tool can I use to check for bad blocks in Linux?

thanks,
007

[voidmain]

fsck. See "man fsck".