Hi,
I just installed SuSE 9.0 on a dual Opteron file server. It has a 4 TB SATA raid attached by Fibre through a Qlogic 2312 HBA. I had issues with the hardware RAID set up (not related to SuSE), originally intending to have a single 3.5TB device (RAID5), but ended up with 2 logical raid devices, which appear to SuSE as /dev/sda (2TB) and /dev/sdb (1.5TB). It takes 20 hours for the RAID to build so I don't want to change that now.
So, in order to combine the space (and allow for future expansion) I set up LVM. I created a volume group that includes both physical volumes, then attempt to create logical volumes. Actually, when creating the volume group there is already a warning that the logical volume size is limited to 2 TB. I have read conflicting information about this. In some places, it says that this is due to the maximum block device size in 2.4 kernels is 2TB, in some places, I read that there is a 2TB limit imposed by 32 bit hardware (but hey, this is 64 bit hardware!). Isn't LVM supposed to get around block device limits? Also, I am using JFS for the file system, so that should not be limiting. When creating the volume group I used 256 MB physical extents (the default is 32) , so that should allow larger logical volumes.
So, I create 2 logical volumes, lvol1 (2TB) and lvol2 (1.5TB), and create a JFS file system on each. In both cases, mkfs.jfs reports the file system created sucessfully. Now, I can mount lvol1, but when I try to mount lvol2, it says "wrong fs type, bad option, bad superblock, or too many mounted filesystems" Well, the mount parameters are the same for both logical volumes (different mount points), and they are both freshly created filesystems, so the only possibility seems is that there are too many mounted filesystems. But, the only other file system that is mounted is the IDE boot disk, which contains the OS, /home, etc. all in a single partition, plus /dev/shm.
Is there a kernel parameter I can tweak to allow me to mount more drive space? 2TB is nothing any more (haha).
Any help will be apreciated.
John Craig