Replaced 2 IDE drives; CentOS only sees 1

JMCraig · 01-06-2010, 05:21 PM

Hi Folks,

I recently had to reload CentOS (5.3) on a box because of an unrelated problem (one of its SATA drives went belly up). I decided to update the old (reliability questionable) 80 GB IDE HDs that had been in it with some new 160's while I was changing stuff around.

Well, the BIOS sees both IDE drives, and so does a bootable Acronis CD I have (boots some Windows version). But CentOS refuses to see both the IDE drives (the new SATA drive dropped in w/o difficulty). I spent some time messing with Master/Slave settings and switching disks around--CentOS only sees the Master disk (whichever physical drive that happened to be at the time--clearly was not a bad drive).

Here's what dmesg has to say for itself in the relevant section:

Code:

Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
NFORCE-MCP55: IDE controller at PCI slot 0000:00:04.0
NFORCE-MCP55: chipset revision 161
NFORCE-MCP55: not 100% native mode: will probe irqs later
NFORCE-MCP55: BIOS didn't set cable bits correctly. Enabling workaround.
NFORCE-MCP55: 0000:00:04.0 (rev a1) UDMA133 controller
    ide0: BM-DMA at 0x2480-0x2487, BIOS settings: hda:DMA, hdb:DMA
    ide1: BM-DMA at 0x2488-0x248f, BIOS settings: hdc:pio, hdd:pio
Probing IDE interface ide0...
ide0: Wait for ready failed before probe !
hda: WDC WD1600AAJB-56R1A0, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Probing IDE interface ide1...
Probing IDE interface ide1...
hda: max request size: 512KiB
hda: 312581808 sectors (160041 MB) w/8192KiB Cache, CHS=19457/255/63, UDMA(133)
hda: cache flushes supported
 hda: hda1 hda2
ide-floppy driver 0.99.newide

lspci says this about the IDE controller:

00:04.0 IDE interface: nVidia Corporation MCP55 IDE (rev a1)

The MB is a Tyan S2912 w/ 2 Dual-core Opterons (with an admittedly older BIOS, but the BIOS sees the drives; and so does the different OS so it seems like it's gotta be something w/ CentOS). The extra space of the 2nd IDE drive was not critical, so I just took it out for the time being (may try a BIOS update later).

Anyone ever encountered this kind of problem before? Could it be as simple as just needing a driver from nVidia?

Thanks.

DaneM · 01-07-2010, 01:12 AM

Hi, JMCraig.

I unfortunately don't have a solution for you, but I can tell you that in every version of Acronis that I know of, the bootable CD IS a customized Linux distribution. That means that in some form, Linux has detected your hardware. While Acronis isn't exactly decent for diagnosing stuff, chances are that if you were to boot up with a Knoppix DVD or Ubuntu Live CD, you will be able to do something like, "lsmod > MyFileOnaUSBDrive" on the working live disc, and then on your dysfunctional CentOS installation/bootable disc. Compare the two and then you'll have a decent idea of what modules need to be loaded. You may want to use the "sort" command to alphabetize the lists of modules for ease of reading.

Hope that helps.

--Dane

strick1226 · 01-07-2010, 09:56 AM

There were a lot of funky BIOS-related issues with Nforce2 chipsets and the IDE controllers not working correctly with kernel 2.6 back ~2006 if I remember correctly. Mainly that stuff was only one of the drives would work at full DMA connection speed, the other would be stuck at DMA/33 or at a CPU-hogging, painfully slow parallel I/O.

While I'm not completely sure just why this never occurred with the original 2x 80 GB IDE hard drives in the box, perhaps it's worth updating the BIOS on the board and seeing if that doesn't fix it? You could always back-flash the BIOS to the current revision if necessary.

Good luck!

Dave_Devnull · 01-07-2010, 11:33 AM

What does the {root} output of 'fdisk -l' say?

JMCraig · 01-07-2010, 11:43 AM

Quote:

Originally Posted by Dave_Devnull

What does the {root} output of 'fdisk -l' say?

Sorry; I thought this was clear, but just to be explicit:

fdisk -l shows the one IDE drive and the two SATA's; the slave IDE doesn't show up at all:

Code:

# fdisk -l

Disk /dev/hda: 160.0 GB, 160041885696 bytes
255 heads, 63 sectors/track, 19457 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hda1   *           1          34      273073+  fd  Linux raid autodetect
/dev/hda2              35       19457   156015247+  8e  Linux LVM

Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          34      273073+  fd  Linux raid autodetect
/dev/sda2              35        4223    33648142+  8e  Linux LVM
/dev/sda3            4224       60801   454462785   8e  Linux LVM

Disk /dev/sdb: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1   *           1          34      273073+  fd  Linux raid autodetect
/dev/sdb2              35        4223    33648142+  8e  Linux LVM
/dev/sdb3            4224       60801   454462785   8e  Linux LVM

Disk /dev/md0: 279 MB, 279511040 bytes
2 heads, 4 sectors/track, 68240 cylinders
Units = cylinders of 8 * 512 = 4096 bytes

Disk /dev/md0 doesn't contain a valid partition table

Note that the various partition types are irrelevant. When I first installed the three new disks (2 IDE and 1 SATA), obviously, there were no partitions on the IDE's and I just wiped everything off the SATA that was still there from the prior setup. The installer couldn't see the slave IDE either.

Haven't had a chance to try lsmod comparison or flashing the BIOS yet.

Dave_Devnull · 01-07-2010, 12:04 PM

Apologies for asking for fdisk -l, it just extends the story being told. Probably stating the obvious, but the new 160gb IDE drives are part of a RAID set, yes? That is the BIOS IDE controller is doing some kind of (software) raid here - or do you have something fancy like an Adaptec, 3Ware or Areca type card?

JMCraig · 01-07-2010, 02:42 PM

Hey Dave, thanks for taking an interest.

But, no, the IDE drives were blank (no partitions of any kind) when the problem first manifested itself. There was/is no BIOS RAID setup (indeed the MB's BIOS does not have any RAID support for IDE drivess anyway--and I'm not using its SATA RAID capability). The RAID partitions you see were set up via CentOS during the install; so, it was applied AFTER the problem of not being able to see the slave IDE drive occurred.

To recap: CentOS could not see the IDE slave drive at all. It never had any partitions built on it nor any BIOS-level RAID setup. I was trying to make that clear above (with my comment about the partition types being irrelevant), but I should have pointed out that the RAID is not BIOS-level, but was set up during the CentOS install (on the 3 disks it could see--of the 4 installed in the box). Therefore, only the master IDE drive's hda1 partion got included in the RAID array along w/ sda1 & sdb1 (setup is RAID 1 w/ a spare).

To conclude: no BIOS RAID involvement.

John

JMCraig · 01-08-2010, 04:23 PM

Turns out the BIOS is actually the latest available already (I was looking at the wrong info when I decided it was an old version)--which is convenient, I guess, 'cause Tyan's site won't let me download the data to reflash it.... Haven't had a chance to do all the fiddling to see if a different driver will work. I did verify that Xubuntu and Ubuntu won't boot from a CD in that box (now I remember why I went with CentOS...).

More later if I make any progress on it.

John