Resize GPT Multipath Disk with Parted on CentOS 6

Online resize of an iSCSI multipath device with GPT partition. Requires no server restart, but the partition needs to be unmounted and not in use.

The Procedure

1. Resize the LUN on a SAN,
2. Rescan devices and reload multipath,
3. multipath -ll show correct size,
4. parted – recreate the partition,
5. partprobe – inform the OS of partition table changes,
6. e2fsck – check the ext4 filesystem before resizing,
7. resize2fs – resize the ext4 filesystem.

Before Resizing the LUN on a SAN

Grab some data. Our multipath device is mpatham and has a size of 20GB. We will extend it by 10GB to 30GB total.

# multipath -ll mpatham
mpatham (36000eb3959dd493f000000000001c3db) dm-77 LEFTHAND,iSCSIDisk
size=20G features='0' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=1 status=active
  |- 93:0:0:0 sdbt 68:112 active ready running
  `- 94:0:0:0 sdbu 68:128 active ready running

# parted /dev/mapper/mpatham print
Model: Linux device-mapper (multipath) (dm)
Disk /dev/mapper/mpatham: 21.5GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      1049kB  21.5GB  21.5GB  ext4         primary

# blockdev --getsz /dev/mapper/mpatham
41943040
# blockdev --getsz /dev/mapper/mpathamp1
41940959

# blockdev --getsz /dev/sdbt
41943040
# blockdev --getsz /dev/sdbt1
41940959

# blockdev --getsz /dev/sdbu
41943040
# blockdev --getsz /dev/sdbu1
41940959

# df -h|egrep "File|mpathamp1|data"
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/mpathamp1
                       20G  142M   20G   1% /data

After Resizing the LUN on a SAN

Parted

You should have parted installed on your CentOS 6 system already. Run the following if it’s not the case:

# yum install parted

Rescan Devices

You can rescan the sdbt and the sdbu devices, or you can rescan all running sessions.

Rescan the devices:

# echo 1 > /sys/block/sdbt/device/rescan
# echo 1 > /sys/block/sdbu/device/rescan

Alternatively, rescan all running sessions:

# iscsiadm -m session --rescan

Note that if you are creating a new device rather than extending an existing one, then you may need to scan the SAN and login to all discovered targets:

# iscsiadm -m discovery -t sendtargets -p netappsan.local:3260
# iscsiadm -m node -L automatic

Reload the multipath service to pick up the new size:

# /etc/init.d/multipathd reload

The multipath device size is 30GB.

# multipath -ll mpatham
mpatham (36000eb3959dd493f000000000001c3db) dm-77 LEFTHAND,iSCSIDisk
size=30G features='0' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=1 status=active
  |- 93:0:0:0 sdbt 68:112 active ready running
  `- 94:0:0:0 sdbu 68:128 active ready running

# blockdev --getsz /dev/mapper/mpatham
62914560
# blockdev --getsz /dev/mapper/mpathamp1
41940959

# blockdev --getsz /dev/sdbt
62914560
# blockdev --getsz /dev/sdbt1
41940959

# blockdev --getsz /dev/sdbu
62914560
# blockdev --getsz /dev/sdbu1
41940959

Recreate the GPT Partition

Make sure the multipath device is not mounted:

# umount /dev/mapper/mpathamp1

Run parted on the multipath device:

# parted /dev/mapper/mpatham
GNU Parted 2.1
Using /dev/mapper/mpatham
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) p
Error: The backup GPT table is not at the end of the disk, as it should be.  This might mean that another operating
system believes the disk is smaller.  Fix, by moving the backup to the end (and removing the old backup)?
Fix/Ignore/Cancel? Fix
Warning: Not all of the space available to /dev/mapper/mpatham appears to be used, you can fix the GPT to use all of the
space (an extra 20971520 blocks) or continue with the current setting?
Fix/Ignore? Fix
Model: Linux device-mapper (multipath) (dm)
Disk /dev/mapper/mpatham: 32.2GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      1049kB  21.5GB  21.5GB  ext4         primary

(parted) rm 1
(parted) unit Kb
(parted) mkpart primary ext4 1049 -1
Warning: You requested a partition from 1049kB to 32212254kB.
The closest location we can manage is 1049kB to 32212237kB.
Is this still acceptable to you?
Yes/No? Yes
(parted) p
Model: Linux device-mapper (multipath) (dm)
Disk /dev/mapper/mpatham: 32212255kB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End         Size        File system  Name     Flags
 1      1049kB  32212238kB  32211189kB  ext4         primary

(parted) q

Reload the Partition Table and Resize the Filesystem

Tell the kernel to reload the partition table. Note that in RHEL 6, partprobe will only trigger the OS to update the partitions on a disk that none of its partitions are in use. If any partition on a disk is in use, partprobe will not trigger the OS to update partitions in the system because it is considered unsafe in some situations.

# partprobe

Check the filesystem before resizing:

# e2fsck -f /dev/mapper/mpathamp1
e2fsck 1.41.12 (17-May-2010)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/mapper/mpathamp1: 11/1310720 files (0.0% non-contiguous), 118545/5242619 blocks

Resize the filesystem:

# resize2fs /dev/mapper/mpathamp1
resize2fs 1.41.12 (17-May-2010)
Resizing the filesystem on /dev/mapper/mpathamp1 to 7864059 (4k) blocks.
The filesystem on /dev/mapper/mpathamp1 is now 7864059 blocks long.

Grab some data:

# blockdev --getsz /dev/mapper/mpatham
62914560
# blockdev --getsz /dev/mapper/mpathamp1
62912479

# blockdev --getsz /dev/sdbt
62914560
# blockdev --getsz /dev/sdbt1
62912479

# blockdev --getsz /dev/sdbu
62914560
# blockdev --getsz /dev/sdbu1
62912479

Finally, mount the partition:

# mount /dev/mapper/mpathamp1 /data/

# df -h|egrep "File|mpathamp1|data"
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/mpathamp1
                       30G  142M   30G   1% /data