On Thu, Sep 15, 2005 at 12:28:01AM +1000, Nick Bryant enlightened us:
On 9/12/05, Joshua Baker-LePain jlb17@duke.edu wrote:
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.
Wow! Odd! RH says 8TB but ext3 FAQ says 4TB.
I wouldn't call it that odd. RH patches their kernels to a fair extent, both for stability and features.
From my personal testing on CentOS 4.1, you can't go over 4TB without
kludging.
I then did a software RAIDO across them, and finally:
mke2fs -b 4096 -j -m 0 -R stride=1024 -T largefile4 /dev/md0
Joshua, thanks for the reply on this. There's something kludgy about having to do softraid across 2 partitions before formatting. It adds a layer of complexity and reduces reliability. Is that the trick RH recommended to go up to 8TB?
Err, it's not a kludge and it's not a trick. Those 2 "disks" are hardware RAID5 arrays from 2 12 port 3ware 9500 cards. I like 3ware's hardware RAID, and those are the biggest (in terms of ports) cards 3ware makes. So, I hook 12 disks up to each card, and the OS sees those as 2 SCSI disks. I then do the software RAID to get 1) speed and 2) one partition to present to the users. Folks (myself included) have been doing this for years.
The one gotcha in this setup (other than not being able to boot from the big RAID5 arrays, since each is >2TiB) is that the version of mdadm shipped with RHEL4 does not support array members bigger than 2TiB. I had to upgrade to an upstream release to get that support.
Just out of interest, and to complicate the matter even more, does anyone know what the upper limit of GFS is?
From what I've been reading, there's an 8TB limit of all GFS file systems in a cluster.
http://www.redhat.com/docs/manuals/csgfs/browse/rh-gfs-en/s1-sysreq-fibredev...
Matt