raid-1 with ide-scsi emulation and mkswap failure

raid-1 with ide-scsi emulation and mkswap failure

[Eric Wood]

I did Warren's RAID-1 howto:
http://togami.com/~warren/guides/remoteraidcrazies/
===========
1.  On this system I raid-1 /dev/hda and /dev/hdc together.

And I'm getting these errors during the copy process.
hdc: dma_intr: error=0x84 { DriveStatusError BadCRC }
hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Come to find out Fedore Core 1 put's the "ide-scsi" in the second ide
channel since I've got a CDRW there (hdd).  Doing a 'rmmod ide-scsi' took
care of all the dma errors and the copy process goes through fine.  So I
removed the "ide-scsi" from grub too.

I assume that ide-scsi on a same channel as a raid-1 member is a big no-no?
Maybe I should be putting raid devices on the same channel now so that dma
is at it's highest setting for both devices.
===========

===========
2. On another system with the same raid-1 setup, I'm getting kernel errors:
 kernel:  Unable to handle kernel NULL pointer dereference at virtual
address 00000015
 kernel:  printing eip:
 kernel: c0159772
 kernel: *pde = 04f30067
 kernel: *pte = 00000000
 kernel: Oops: 0000, etc. ,etc.

So I decided to redo the swap partition (which is ARRAY /dev/md3
devices=/dev/hdd3,/dev/hda3):

# swapoff -a
# mdadm --stop --scan
# sync; mkswap /dev/md3
mkswap: error: swap area needs to be at least 40kB
Usage: mkswap [-c] [-v0|-v1] [-pPAGESZ] /dev/name [blocks]

What I don't understand is that the 3rd partion of both drives which are
identical have 522112 blocks available:
# sfdisk -l /dev/hda

Disk /dev/hda: 4865 cylinders, 255 heads, 63 sectors/track
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

  Device Boot Start     End   #cyls    #blocks   Id  System
/dev/hda1   *      0+     12      13-    104391   fd  Linux raid autodetect
/dev/hda2         13     904     892    7164990   fd  Linux raid autodetect
/dev/hda3        905     969      65     522112+  fd  Linux raid autodetect
/dev/hda4        970    4864    3895   31286587+  fd  Linux raid autodetect

# sfdisk -l /dev/hdd

Disk /dev/hdd: 4865 cylinders, 255 heads, 63 sectors/track
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

  Device Boot Start     End   #cyls    #blocks   Id  System
/dev/hdd1   *      0+     12      13-    104391   fd  Linux raid autodetect
/dev/hdd2         13     904     892    7164990   fd  Linux raid autodetect
/dev/hdd3        905     969      65     522112+  fd  Linux raid autodetect
/dev/hdd4        970    4864    3895   31286587+  fd  Linux raid autodetect

I'll trying removing the swap line out of /etc/fstab and reboot, and try and
add it back but I have a feeling that won't work.  Any other ideas?
===========


-Eric Wood

Re: raid-1 with ide-scsi emulation and mkswap failure

[Tim Small]

Eric Wood wrote:

>I did Warren's RAID-1 howto:
>http://togami.com/~warren/guides/remoteraidcrazies/
>===========
>1.  On this system I raid-1 /dev/hda and /dev/hdc together.
>
>And I'm getting these errors during the copy process.
>hdc: dma_intr: error=0x84 { DriveStatusError BadCRC }
>hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
>  
>
If you haven't already, then it may be worth checking the output of 
smartctl -a /dev/hdc  - just to satisfy yourself that the drive is 
feeling happy with itself...

http://smartmontools.sourceforge.net/

Tim.

Re: raid-1 with ide-scsi emulation and mkswap failure

[Eric Wood]

Okay, Catalin showed me my goof on the mkswap problem.  So maybe redoing the
swap will keep the kernel from panicing any more.
-Eric Wood

----- Original Message ----- 
From: "Eric Wood"
> # swapoff -a
> # mdadm --stop --scan

----- Original Message ----- 
From: "Catalin BOIE"> > # swapoff -a
> > # mdadm --stop --scan
>
> Why did you stopped the array?
> It must be started when you do mkswap /dev/md3.

Ah, your absolutely right! Duh!!

Thanks,
-eric wood

Re: raid-1 with ide-scsi emulation and mkswap failure

[Daniel Pittman]

On Thu, 15 Apr 2004, Eric Wood wrote:
> I did Warren's RAID-1 howto:
> http://togami.com/~warren/guides/remoteraidcrazies/
> ===========
> 1.  On this system I raid-1 /dev/hda and /dev/hdc together.
> 
> And I'm getting these errors during the copy process.
> hdc: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Others have pointed out that ide-scsi shouldn't be the cause of this,
but no one seems to have noted the meaning of these errors.

Specifically, the BadCRC error says that the data transfer across the
IDE bus was corrupted, usually a sign that your data cable is too long.

You may want to look into that if this is more than a transient problem.

    Daniel

-- 
I have always detested the belief that sex is the chief bond
between man and woman. Friendship is far more human.
        -- Agnes Smedley