RHEL: Extending a multipath LUN

# Tested on RHEL 6


# This procedure may be carried out to make visible the new size of a LUN that is already
# presented and used by the OS and that has been extended at SAN storage level.

# Whenever possible I recommend to create a new LUN at SAN level instead of doing an
# extension. In some situations we may experience some trouble when trying to recognize
# the new size of a disk that has been extended so a reboot may be necessary.


# Find the paths to the LUN:

multipath -ll
[...]
   my_lun_01 (200255c3a11080003) dm-38 NEC,DISK ARRAY
   size=120G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
   |-+- policy='round-robin 0' prio=50 status=active
   | |- 1:0:2:0 sdm 65:32 active ready running
   | `- 2:0:2:0 sdo 65:64 active ready running
   `-+- policy='round-robin 0' prio=10 status=enabled
     |- 1:0:3:0 sdn 65:16 active ready running
     `- 2:0:3:0 sdp 65:48 active ready running



# Rescan the paths:

echo 1 > /sys/block/sdm/device/rescan
echo 1 > /sys/block/sdo/device/rescan
echo 1 > /sys/block/sdn/device/rescan
echo 1 > /sys/block/sdp/device/rescan




# Resize the multipath device:

multipathd -k'resize map my_lun_01'
   ok

# or, in interactive mode:
# multipathd -k
# multipathd> resize map my_lun_01
# ok
# multipathd> exit



# Check

multipath -ll
[...]
   my_lun_01 (200255c3a11080003) dm-38 NEC,DISK ARRAY
   size=120G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
   |-+- policy='round-robin 0' prio=50 status=active
   | |- 1:0:2:0 sdm 65:32 active ready running
   | `- 2:0:2:0 sdo 65:64 active ready running
   `-+- policy='round-robin 0' prio=10 status=enabled
     |- 1:0:3:0 sdn 65:16 active ready running
     `- 2:0:3:0 sdp 65:48 active ready running




# At this point, physical volume size shown by 'pvdisplay' and 'fdisk' should be different.


# If existing physical volume was created directly on the whole disk, without partition,
# a 'pvresize' should be enough for the new size to be taken into account

pvresize /dev/$SD

# If, on the other hand, disk is already partitioned, this is, we are using devices in the
# form /dev/sdx1, /dev/sdx2, we have to create a new partition with 'fdisk' tool

fdisk /dev/$SD

   The number of cylinders for this disk is set to 2480.
   There is nothing wrong with that, but this is larger than 1024,
   and could in certain setups cause problems with:
   1) software that runs at boot time (e.g., old versions of LILO)
   2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) 
 
   Command (m for help): p
 
   Disk /dev/sdc: 20.4 GB, 20401094656 bytes
   255 heads, 63 sectors/track, 2480 cylinders
   Units = cylinders of 16065 * 512 = 8225280 bytes   
 
   Device Boot      Start         End      Blocks   Id  System
   /dev/sdc1   *           1          33      265041   83  Linux
   /dev/sdc2              34        1958    15462562+  8e  Linux LVM


# Let's create a new partition. In our case, we'll create partition # 3
# (primary Linux LVM partition). For the first and last cylinder usually default values
# will be ok; if not, choose carefully the beginning and the end of the new partition to
# avoid

   Command (m for help): n
   Command action   
   e   extended   
   p   primary partition (1-4)
   p
   Partition number (1-4): 3
   First cylinder (1959-2480, default 1959):
   Using default value 1959
   Last cylinder or +size or +sizeM or +sizeK (1959-2480, default 2480):
   Using default value 2480 
 
   Command (m for help): t
   Partition number (1-4): 3
   Hex code (type L to list codes): 8e
   Changed system type of partition 3 to 8e (Linux LVM) 
 
   Command (m for help): w
   The partition table has been altered!
 
   Calling ioctl() to re-read partition table.
 
   WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
   The kernel still uses the old table.
   The new table will be used at the next reboot.
   Syncing disks.

# Rescan disks

partprobe -s

# It may be that we have an error like following on (usually on RHEL 6):

Warning: WARNING: the kernel failed to re-read the partition table on /dev/sdc (Device or resource busy).
As a result, it may not reflect all of your changes until after reboot.

# The use following command instead:

partx -a /dev/$SD
 

# New partition is ready to be used ('pvcreate', etc)

brw-r----- 1 root disk 8, 33 Nov  3 15:15 /dev/sdc1

0 (0)
Article Rating (No Votes)
Rate this article
Attachments
There are no attachments for this article.
Comments
There are no comments for this article. Be the first to post a comment.
Full Name
Email Address
Security Code Security Code
Related Articles RSS Feed
Installing and Configuring an OCFS2 Clustered File System
Viewed 6102 times since Sat, Jun 2, 2018
Zabijanie wszystkich procesów użytkownika
Viewed 2865 times since Thu, May 24, 2018
12 Tcpdump Commands – A Network Sniffer Tool
Viewed 8924 times since Fri, Jul 27, 2018
RHEL: iSCSI target/initiator configuration on RHEL6
Viewed 9029 times since Sat, Jun 2, 2018
30 Handy Bash Shell Aliases For Linux / Unix / MacOS
Viewed 5105 times since Thu, Feb 11, 2021
debian Install a newer kernel in Debian 9 (stretch) stable
Viewed 1962 times since Sun, Sep 23, 2018
How to find the largest files and directories in Linux?
Viewed 3589 times since Sun, May 20, 2018
RHEL: Reinstalling Boot Loader on the Master Boot Record (MBR)
Viewed 3603 times since Sun, May 27, 2018
How log rotation works with logrotate
Viewed 5085 times since Fri, Nov 30, 2018
how to list all hard disks in linux from command line
Viewed 3838 times since Mon, Jan 28, 2019