For some reason everytime I try to format a md device on one of my servers it just halts the whole machine. This is how far it gets:


[root@extsrv ~]# mkfs.ext3 /dev/md0
mke2fs 1.40.8 (13-Mar-2008)
Warning: 256-byte inodes not usable on older systems
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
61054976 inodes, 244190624 blocks
12209531 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=4294967296
7453 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
102400000, 214990848

Writing inode tables: 142/7453




I tried to create a partition on the md device but I get an error (it's always done this on all servers I've setup) so I just ignored it and formatted the drive directly, I've done this many times no issues till now.

I also get this error at startup, could it be bad ram?

PCI: Error while updating region 0000:00:0d,1.1 (e2040000 != (missed the rest - had to be fast with the camera)

I see nobody's jumping in so maybe I can offer some ideas if not The Answer. Let's see if I understand the situation.
- This is happening on several different machines.
- All of the machines are running the same distribution and version of Linux.

This reminds me of when True Crypt released version 5.x. It caused exactly the same problem. The True Crypt people blamed the kernel. They also claimed that the problem was fixed in v2.6.24 of the kernel. I never tested to see if it was true.

You may be asking why the computer processes eventually freeze. If the physical disk drive that you are formatting also contains one or more partitions that are mounted by Linux then this is what happens.
- Your mkfs process issues a blocking i/o to write to the disk. For some reason that hangs. Then as each process attempts to read or write to the same physical disk, regardless of what partition, Linux puts those i/o requests behind the one from mkfs that is hung. Thus as each process tries to access the disk it will be put in a permanent wait state.

What Linux kernel version are you running? If you are running a version prior to 2.6.24 then you may want to upgrade or just stop using mkfs on the running system. You could boot a live CD and format your partitions that way. That would be a good diagnostic technique to test my blocking i/o theory.

I'm doing memtest now and going to leave it overnight.

(only doing this on one machine)

I'm running whatever kernel comes with FC9. I was thinking maybe the system was getting confused as the disks were previously used in a raid5 and now I'm using them in a raid 1 so I did a dd, it crashed durring the dd too but this time I got this error:

tomorrow I'll try the live cd route. Guess I should probably use another one then the FC9 one to rule out that kernel. I think I may have an ubuntu disc around.

ok I doubt it's the drive. Did self test using smartctl and they came clean. I also tried to format them individually with ext3, they BOTH freeze. This makes no sense. It even does it if I use a boot CD. Anything that writes to it basically locks up the entire system. That machine was my old server, so it can't be the controller either, or I would of had problems before.


Edit: I can confirm it's not the drives. I tested both in another system and they format fine. I'm really at loss here. I tried plugging them into the other sata ports (I have 2 red ones and two black ones) but it would only see 1 drive. Is my motherboard screwed? I really hope not, as it's an older one and will be very hard to replace, I really can't afford to build a whole new system. I already spent enough money on this project, but now it's falling in the water. Would it work if I got a pci sata controller card?