Red Hat Bugzilla – Bug 115889
(dmraid) Promise 20378 RAID volume not recognized
Last modified: 2007-11-30 17:10:36 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6b)
Description of problem:
When booting with the fedora core 2 test 1 CD, my RAID-0 on my Promise
20378 SATA controller is not recognized. The motherboard is a ASUS
P4C800-E. I had prevously set up the RAID-0 with fasttrak
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Insert Fedora Core 2 test 1 CD and boot
Actual Results: Drive not recognized
Expected Results: Drive recognized
There's no support in the kernel for this in 2.6
I have an MSI K8T NEO motherboard with promise raid controller as
well, ATHLON 64 3000+, and am installing FEDORA CORE 2 production
version. A similar problem occurs. On initial install, the promise
chip is recognized but install cannot see the 2 hard drives in a
hardware RAID 0 as a single drive. Even after I initialize the
drives, and the install seems to complete, the machine freezes on
GRUB at boot.
Ditto my ASUS K8V Deluxe, Athlon 64 3200+. Using 2-disk Parallel ATA
Raid 0 on the Promise 20378 controller. Trying to make WinXP/FC dual
boot system. WinXP sees raid array and boots fine. Trying to install
FC2 64-bit, controller is recognized as an SATA device, but defined
parallel ATA raid 0 array is not seen at all.
Still no support on FC3 final.
My guess is that the problem lies with anaconda as I'm getting
Interrupt #11 errors (BIOS equipment enumeration).
Anaconda's quiering the BIOS, it's responding, but doesn't know whom
to hand off the interrupt to the proper driver.
You may use 'dmraid' to access your Promise RAID, but that is not
supported at install time by FC2/FC3.
Until the installer supports dmraid, you will need to use the 'md'
software RAID that the installer -does- support.
Is there any hope of making a kernel module so that when the installer
boots, a "linux dd" command could then be used to prompt for a driver
diskette so that the install could see the RAID volume?
The stock ata driver is choking on the extra
interrupt #11s generated -- even as FC3 starts the xserver, the
interrupt errors keep on coming, even when starting FC3 with "linux dd
acpi=off" and loading the vendor-supplied driver.
Jeff Garzik's libata patch at
solves this problem, as per my troubles noted on
But it is verry difficult to create a boot disk that works (for me it was, anyhow).
Can this patch be included in a separate boot.iso somewhere, for the flood of
people who are going to hit this before FC4 comes out? I could make the boot.iso
(now that I've climbed the learning curve), but I have nowhere to upload it.
It would be great if they would merge the dmraid features into the
kernel (which is the place where they have to go). I have waited for
years to be able to boot linux from my PDC30378-'Raid'
dmraid (userspace) uses device mapper (kernel), which is the proper
separation of components. Nothing else needs to be in the kernel.
The actual RAID1 driver, for example, is dm-mirror (kernel), but it
requires the userspace component to set it up.
This bug in no way relates to any Promise hardware/driver issues.
Please do not add sata_promise-related hardware/driver bugs to this issue.
The rationale is that the faulty libata code was interfering with the Promise
module's ability to communicate with the hardware.
Fedora Core 3 is now maintained by the Fedora Legacy project for security
updates only. If this problem is a security issue, please reopen and
reassign to the Fedora Legacy product. If it is not a security issue and
hasn't been resolved in the current FC5 updates or in the FC6 test
release, reopen and change the version to match.
The bug still exists in Core 5.
We've got anaconda support for dmraid in FC5 for a while now.
Is the previous comment based on a test result ?
If so, please attach the output metadata files created with 'dmraid -rD',
tar'ed/bzip2*ed in bz2">firstname.lastname@example.org for reference.
Moving to FC5 as per comment #17.
No followup since a while.