ext3 +2TB fs
alazarev at itg.uiuc.edu
Fri Feb 25 22:36:33 UTC 2005
ah, found the error logs in a previous email i never sent:
Feb 25 11:23:28 xxx kernel: sda: sda1
Feb 25 11:23:30 xxx kernel: SCSI device sda: 2927173632 512-byte hdwr
sectors (1498713 MB)
Feb 25 11:23:30 xxx kernel: SCSI device sda: drive cache: write back
Feb 25 11:23:30 xxx kernel: sda: sda1
Feb 25 11:28:25 xxx kernel: EXT3-fs error (device sda1):
ext3_check_descriptors: Inode table for group 5642 not in group (block
Feb 25 11:28:25 xxx kernel: EXT3-fs: group descriptors corrupted !
Feb 25 11:28:34 xxx kernel: EXT3-fs error (device sdb1):
ext3_check_descriptors: Inode table for group 826 not in group (block
Feb 25 11:28:34 xxx kernel: EXT3-fs: group descriptors corrupted !
At that point I can fsck and fix the fs, but as soon as I try to mount the
problem happens again.
that is the error i get on a RHEL4-AS 2.6.9-5.0.3smp kernel using parted
1.6.21 and mkfs command from below.
any ideas? thanks!
On Fri, 25 Feb 2005, Alexander Lazarevich wrote:
> I've got a 3.3TB ext3 on a FC3 64-bit system, running kernel
> 2.6.10-1.766FC3smp. I create the partition with parted 1.6.21, and I make the
> fs via:
> mkfs.ext3 -m1 -b 4096 -T largefile4 /dev/sda1
> Works fine. bonnie++ running on it multiple times for days on end, no
> However, I do the exact same setup on a RHEL4-AS i686 system, 32-bit, and the
> fs is totally hosed, get all kinds of errors with the filesystem, just trying
> to mount it doesn't work. Crud, I blew away the system, so I don't have the
> error logs, but it was an error about the group descriptors being corrupted.
> I ran fsck but it couldn't fix it. That kernel is 2.6.9-5.0.3smp. I also
> compile with the CONFIG_LDB flag which enables large block device support.
> I'm told that should be all I have to do. But it just doesn't work. Not at
> Any ideas? I'm stuck running FC3 now unless I can get some quick fix from you
> Ext3-users mailing list
> Ext3-users at redhat.com
More information about the Ext3-users