[CentOS] mkfs.ext3 on a 9TB volume

[Joshua Baker-LePain]

On Sun, 11 Sep 2005 at 6:41pm, Francois Caen wrote

> On 9/11/05, Joshua Baker-LePain <jlb17 at duke.edu> wrote:
> > Having hit a similar issue (big FS, I wanted XFS, but needed to run centos
> > 4), I just went ahead and stuck with ext3.  My FS is 5.5TiB -- a software
> > RAID0 across 2 3w-9xxx arrays.  I had no issues formatting it and have had
> > no issues in testing or production with it.  So, it can be done.  Perhaps
> > the bugs you're hitting are in the FC driver layer?
> 
> ext3 had a 4TB limit:
> http://batleth.sapienti-sat.org/projects/FAQs/ext3-faq.html)
> which I didn't know when I started this thread.

As I mentioned, I'm running centos-4, which, as we all know, is based off 
RHEL 4.  If you go to <http://www.redhat.com/software/rhel/features/>, 
they explicitly state that they support ext3 FSs up to 8TB.

> I found it the hard way, through testing.
> There are ways to force past that limit (mkpartfs ext2 in parted, then
> tune2fs -j), but the resulting filesystem is totally unstable.
> 
> Joshua, how the heck did you format your 5.5TB in ext3? You 100% sure
> it's not mounted as ext2?

To answer the 2nd question:
[jlb@$HOST ~]$ df -h
Filesystem            Size  Used Avail Use% Mounted on
.
.
/dev/md0              5.5T  634G  4.9T  12% /nefs
[jlb@$HOST ~]$ mount
.
.
/dev/md0 on /nefs type ext3 (rw)

As to the first, I created the FS as simply as possible.  /dev/sdb and 
/dev/sdc both look like this:

(parted) print                                                            
Disk geometry for /dev/sdb: 0.000-2860920.000 megabytes
Disk label type: gpt
Minor    Start       End     Filesystem  Name                  Flags
1          0.017 2860919.983                                    

I then did a software RAIDO across them, and finally:

mke2fs -b 4096 -j -m 0 -R stride=1024 -T largefile4 /dev/md0

-- 
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University