[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: Using kickstart on large disks

Thanks everyone for your replies. See my comments below:

> (Shabazian, Chip) wrote:
>> Why not just create the /db volume in the %post?

Well, firstly I wanted to use the "logvol" option because it does a lot of work for you, but the main problem with a %post approach is that one of the RPMs that gets installed puts a lot of stuff into /db.

I actually use a hack right now where that RPM uses lvextend & resize2fs to expand the volume to the full 16TB before installing stuff into it.

> (Digimer) wrote:
>> That would be a work-around to Arun's problem without addressing
>> the  underlying issue. With single drives being 2TB now, it's an
>> issue that  deserves direct attention, I would think.

Thanks Digi, and agreed. I currently use the ugly workaround I mentioned above and it works fine, but the fix to anaconda seems really easy to me. More importantly, I feel that this is something that other folks will run into and a fix would save them a lot of headache.

> (Pablo Iranzo G?mez) wrote:
>> http://www.redhat.com/rhel/compare/
>> See the section related to capabilities/limits, specially when it
>> says that for EL5 limits is 8TB, but for 5.1 is 16TB, for bigger FS
>> you'll need gfs. It's usual to have gfs in 'solo' mode for getting
>> better performance for larger storage capabilities.

Thanks Pablo. I will look at GFS but my needs are fairly simple and don't really necessitate a cluster file system. With my current lvextend workaround described above, Ext4 is doing a great job with the 16TB partition. I just wish that anaconda/kickstart wouldn't limit Ext4 partitions to 8T.

> (Steffen Maier) wrote:
>> I think a discussion on this has at least started on anaconda-devel:
>> https://www.redhat.com/archives/anaconda-devel-list/2010-January/msg00291.html

Yeap that is exactly my problem, except that I'm using Ext4 and not XFS. I just wonder if the current 8TB limit is intentional (I'm no filesystem expert) or just an oversight when the Ext4 FS type was extended from Ext2.



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]