[Linux-cluster] DLM Problem

[Gordan Bobic]

isplist at logicore.net wrote:
> Something else going on here.
> 
> I see the device using fdisk;
> 
> Disk /dev/sda doesn't contain a valid partition table
> You have new mail in /var/spool/mail/root
> 
> I then use fdisk to create a partition;
> 
> Disk /dev/sda: 774.2 GB, 774207700992 bytes
> 255 heads, 63 sectors/track, 94125 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
> 
>    Device Boot      Start         End      Blocks   Id  System
> /dev/sda1               1       94125   756059031   83  Linux
> 
> Looks fine right.
> 
> Looking on the net, I can't find anything on gfs_mkfs one single drive so have 
> to make assumptions for now. 
> 
> Assumptions are, lock mode, cluster name, FS name, journals and device.
> 
> #gfs_mkfs -p lock_dlm -t compweb:web -j 8 /dev/sda
> 
> Seems to go fine;
> 
> Device:                    /dev/sda
> Blocksize:                 4096
> Filesystem Size:           188738976
> Journals:                  8
> Resource Groups:           2882
> Locking Protocol:          lock_dlm
> Lock Table:                compweb:web
> 
> Syncing...
> All Done
> 
> Do an fdisk again;
> 
> Disk /dev/sda: 774.2 GB, 774207700992 bytes
> 255 heads, 63 sectors/track, 94125 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
> 
> Disk /dev/sda doesn't contain a valid partition table
> 
> It's all gone. Guessing I need to gfs_mkfs some other way?

You're going to hit yourself when I say this, but you made the GFS 
partition on the raw drive (sda rather than sda1), not on a partition. 
There is nothing wrong with making it on the raw disk, mind you, that's 
what I always do when using SAN volumes.

Gordan