Linux LVM Notes

For doing the same in Linux. Note that I usually work with multipathing so the notes below reflect that. Both Powerpath and native Linux multi path are included on the notes. # Create a LVM VG (Linux) # add disk masking & zoning (storage & SAN switch) for i in 0 1 2 3 <...>;do echo "- - -">/sys/class/scsi_host/host$i/scan;done # check for new disks multipath -v2 # -v2 just for additional info, and this just for multipath if you're using powerpath see below powermt config # this is for power path # if you're not using powerpath or multipath proceed using the sdNN device as PV fdisk /dev/mapper/mpathz # create a new partition # this isn't required, LVM will handle the lun even if it's not partitioned but then # fdisk -l will not be so handy # if using powerpath fdisk command will be fdisk /dev/emcpowerz # and for non multipathed luns: fdisk /dev/sdz # your PV device will be /dev/mapper/mpathzp1, /dev/emcpowerz1 or /dev/sdz1 pvcreate /dev/mapper/mpathzp1 # or pvcreate /dev/emcpowerz1 # or pvcreate /dev/sdz1 # if needed create additional physical disks # a volume group is made of one or more physical disks vgcreate vgname /dev/mapper/mpathzp1 # additional disks physical disks may be specified # specify the PV created before # if adding several luns specify them all vgcreate vgname /dev/mapper/mapthzp1 /dev/mapper/mpathyp1 # now create a logical volume lvcreate -l 400 /dev/vgname /dev/vgname/lvol0 # this creates a logical volume '/dev/vgname/lvol0' with 10GB inside volume group 'vgname' # (provided enough space is available lvcreate -l 400 /dev/vgname # creates a logical volume lvolN (N is an ordinal automatically assigned) on volume group # vgname mkfs -t ext4 /dev/vgname/lvol0 # to format logical volume /dev/vgname/lvol0 with ext4 (RH 6, on RH5 should be ext3) /dev/vgname/lvol0 /mountdir ext4 defaults 0 0 # add this line to /etc/fstab so that the filesystem on /dev/vgname/lvol0 will be # mounted automatically on each reboot mount -a # mount all filesystems specified on /etc/fstab which aren't mounted or mount -t ext4 /dev/vgname/lvol0 /mountdir # to mount the new filesystem on a specified directory # Resize a filesystem (online when extending - to shrink it should be unmounted) vgdisplay vgname # check lines 'PE Size' and 'Free PE', it's the space available vgextend /dev/vgname /dev/mapper/mpathaap1s # if needed add more physical volumes to the volume group lvextend -l 800 /dev/vgname/lvol0 # assuming that pe size is 256MB, extend /dev/vgname/lvol0 to 20 GB resize2fs /dev/vgname/lvol0 # resizes the filesystem mounted on /mountdir to 20 GB

HPUX LVM notes

Quite dumb to start this with HPUX, but anyway, there it goes: # Create a LVM VG (HPUX) # add disk masking & zoning (storage & SAN switch) ioscan -fnNCdisk # check for new disks # you'll need the disk device '/dev/disk/diskNNN' ex: /dev/disk/disk175 pvcreate /dev/rdisk/disk175 # note that on pvcreate we use the raw (character) disk device # if needed create additional physical disks # a volume group is made of one or more physical disks vgcreate -e 64000 -l 64 -s 256 -g PVGSTO vgname /dev/disk/disk175 # this creates a volume group with 64000 maximum extents of 256 MB and 64 logical volumes # physical volume group will be named PVGSTO (this is useful when multiple storage tiers # are used in the same volume group # additional disks physical disks may be specified lvcreate -l 400 /dev/vgname /dev/vgname/lvol0 # this creates a logical volume '/dev/vgname/lvol0' with 10GB inside volume group 'vgname' # (provided enough space is available) lvcreate -l 400 /dev/vgname # creates a logical volume lvolN (N is an ordinal automatically assigned) on volume group # vgname newfs -F vxfs -o largefiles /dev/vgname/rlvol0 # to format logical volume /dev/vgname/lvol0 with vxfs - note 'rvol0': the character # device should be specified /dev/vgname/lvol0 /mountdir vxfs delaylog 0 0 # add this line to /etc/fstab so that the filesystem on /dev/vgname/lvol0 will be # mounted automatically on each reboot mount -a # mount all filesystems specified on /etc/fstab which aren't mounted or mount -t vxfs /dev/vgname/lvol0 /mountdir # to mount the new filesystem on a specified directory # Resize a filesystem (online when extending - to shrink it should be unmounted) vgdisplay vgname # check lines 'PE Size' and 'Free PE', it's the space available vgextend -g PVGSTO /dev/vgname /dev/disk/disk176 # if needed add more physical volumes to the volume group lvextend -l 800 /dev/vgname/lvol0 # assuming that pe size is 256MB, extend /dev/vgname/lvol0 to 20 GB lvextend -l 800 /dev/vgname/lvol0 /dev/PVGSTO1 # extends /dev/vgname/lvol0 to 20GB using physical volume group 'PVGSTO1' fsadm -b 20480M /mountdir # resizes the filesystem mounted on /mountdir to 20 GB Enjoy!