I have RHEL 5 server that I add a new SAN LUN to. I'm also using EMC's powerpath to manage multipathing. The new disk showed up after re-scanning the SCSI bus then running the 'powermt config' command. I was trying to configure the new storage for an LVM but on device /dev/emcpowerai but I inadvertantly ran fdisk against /dev/emcpowera1, the partition that houses /boot. So now if I fdisk -l to list the partitions on the disk for /emc/emcpowera I get:

And if I run fdisk -l against /dev/emcpowera1 I get this:

So it looks like I've created a "sub" partition on /dev/emcpowera1.

Obviously, I'm scared to death that I've screwed up my boot partition. I have not yet rebooted and I'm hoping someone can help me back out of the mess I've gotten into. Any suggestions?

First, see if you have overwritten the superblock:

If it says you have a primary at block 1, then you did not. In that case you can umount and run the appropriate type of fsck for your filesystem: ext2, ext3 or ext4. If you have some other FS, then ignore me, since I'm less familiar with those.

Once fsck passes, you can remount and make sure your boot files are still there.

Thank you, looks like things may not be as bad as I feared:

The primary superblock is at 1. I also have a test system that I used to perform the same steps on, recreating my predicament. I created another partition on the boot partition. But before rebooting I went back into fdisk and deleted the partition. I then rebooted the box and it came up fine. But the box I'm concerned about is a production system, so I'm waiting for input from Redhat as well as this forum. But I'm thinking I may be okay just deleting the partition.

You should be OK. Creating a partition table affected only the first 512-byte block in the partition. The primary ext2/3/4 superblock is the 2nd 1K block, so you never touched it.