PDA

View Full Version : Dell PE 1950 won't boot Fedora 14 i386: HDIO_GET_IDENTITY failed and udev errors


rovitotv
6th December 2010, 10:57 PM
We have a few older Dell 1950 PowerEdge servers II that we would like to install Fedora 14 on. The install goes perfect but then Fedora won't boot. Here are the error messages I get:
HDIO_GET_IDENTITY failed for ‘/dev/sda’
Welcome to Fedora
Starting udev: ata1.00: exception emask 0x00 SACT 0x00 SErr 0x00 action 0x6 frozen
(note very long wait time to get to this point several minutes)
ATA1.00: status { DRDY 3 }
Udevd: worker [447] unexpectedly returned with status 0x0100
Udevd worker [447] failed while handling ‘/devices/pci0000:0010000:1f.0’
Udevd: worker [447] unexpectedly returned with status 0x0100
Udevd worker [500] failed while handling ‘/devices/virtual/block/loop0’

This same machine and hard disk has no problems booting/running CentOS 5.5, I even tried different servers and hard disks. So I don’t think it is the hardware. I am at a loss, any ideas? I would really like to use Fedora for the new python and new glibc.


Oddly enough RHEL version 6 installs and boots. I still get the HDIO_GET_IDENTITY failed but no udev errors. RHEL 6 has up to date python and glibc so this gets me started on this project but I would of liked to have used Fedora because it is free.

Thanks for the support.

DBelton
7th December 2010, 12:03 AM
I'll be willing to bet that Fedora 13 will install and run fine as well. I have noticed (and have several different posts on it in here) that I am getting those same errors on a Fedora 14 install that occur randomly on 12 different hard drives, yet if I switch back to Fedora 13, they never occur.

also, just out of curiosity... Do you have a disk utility that can read the SMART table from the drive? I am wondering if yours is now showing "Pending Reallocated Sectors" (which will be cleared when those sectors are successfully read/written)

When that particular error you show above appears on mine, it is when I have read errors on sectors that happen to hold the partition table. Sometimes, I can get it back to working correctly just by doing a complete read on the drive, sometimes, I have to run something like HDAT2 that will do a read/write/read/verify. Occasionally, it will actually have some corrupted data, and I must then run testdisk to recover the partition information and then write it back to the drive.

It is getting to be completely frustrating for me. As I mentioned elsewhere, this has been an ongoing thing for me on F14... Happening on 12 different drives, and I just can't believe that 12 drives are suddenly going bad without a problem under F13 or Windows.

I have thought several times that possibly I am having an intermittent memory error and it just happens to hit the "right" combination of things running on F14, but all memory tests I have done show no problems, as have the numerous surface scans I have done on the hard drives.