Concepts
--------
PHYSICAL VOLUME (PV):
physical devices, local disk or luns
|
|
VOLUME GROUP (VG):
groups of pvs
|
|
LOGICAL VOLUME (LV):
partitions inside vg
|
|
FILESYSTEM (FS):
ext2/3/4,DOS,swap,etc..
Paritions
---------
/proc/partitions ##contains list of known partitions to kernel
|
Listing/Viewing
|
fdisk
fdisk -l ##lists available devices together with
partition tables
fdisk -l | grep
sd ##filters scsi devices
cfdisk
## to lists
partition codes
sfdisk -T
|
|
Modifying
|
## rescans the
whole disk (useful in resizing VMDK disks)
echo 1 >
/sys/block/sd
pvresize
/dev/sdc
partprobe
partx -a
|
|
Debugging
|
dumpe2fs
e2fsck
|
Physical Volumes
----------------
examples of PVs:
/dev/sdb
/dev/sdb[12]
Commands:
|
Displayng
|
## displays info
reg physical volumes
pvdisplay --> more detailed output
pvs --> quick view
pvscan
pvdisplay
--ignorelockingfailure --> use
incase of r/o filesystem
note: you need a
physical device or a partition from a physical device to create a physical
volume
here are the rest
of pv commands:
[mvineza@secutil07
~]$ ls /sbin/pv*
/sbin/pvchange /sbin/pvcreate /sbin/pvmove /sbin/pvresize /sbin/pvscan
/sbin/pvck /sbin/pvdisplay /sbin/pvremove /sbin/pvs
[mvineza@secutil07
~]$
|
|
Creating
|
## creates and
display physical volume/s
pvcreate
## forcely
initialize/reinitialize a physical device (for previously used devices)
pvcreate -ff
|
|
Removing
|
## forcely removes
a physical volume from a volume group
pvremove
|
Volume Groups
-------------
commands:
|
Displaying
|
## displays vg info
vgdisplay
vgdisplay
vgdisplay -v
vgscan --> scan
all disks for volume groups and rebuild caches
vgs --> quick
view
here are list of vg
commands:
$ ls /sbin/vg*
/sbin/vgcfgbackup /sbin/vgck /sbin/vgdisplay /sbin/vgimport /sbin/vgmknodes /sbin/vgrename /sbin/vgsplit
/sbin/vgcfgrestore /sbin/vgconvert /sbin/vgexport /sbin/vgimportclone /sbin/vgreduce /sbin/vgs
/sbin/vgchange /sbin/vgcreate /sbin/vgextend /sbin/vgmerge /sbin/vgremove /sbin/vgscan
$
|
|
Creating/Modifying
|
## creates vg
vgcreate
vgcreate -c n
vgcreate -s
vgcreate -L size
## increase vg
vgextend
## activates a
volume group
vgchange -a y
|
|
Deleting
|
## decrease vg
vgreduce
vgreduce
|
Logical Volumes
---------------
|
Displaying
|
## displays info on
logical volumes
lvdisplay
/dev/
lvscan
lvs
here is a list of
lv commands:
$ ls /sbin/lv*
/sbin/lvchange /sbin/lvdisplay /sbin/lvmchange /sbin/lvmdump /sbin/lvmsar /sbin/lvrename /sbin/lvscan
/sbin/lvconvert /sbin/lvextend /sbin/lvmconf /sbin/lvmetad /sbin/lvreduce /sbin/lvresize
/sbin/lvcreate /sbin/lvm /sbin/lvmdiskscan /sbin/lvmsadc /sbin/lvremove /sbin/lvs
$
|
|
Creating
|
## basic way of
creating an LV
lvcreate [options]
## if you want to
explicitly define its characteristics, use this
lvcreate -n
## if you want to
consume all the space in the volume group
lvcreate -n
|
|
Resizing
|
## increases size
of a logical volume
lvextend -L
+
lvextend -L
+
lvextend -L
lvextend -L +50g
/dev/VG1/LV1 --> for redhat 5.1
lvextend -l
+100%FREE /dev/mapper/
## increasing by pv
(uses whole pv)
lvextend
/dev/mapper/vg_root-lv_root /dev/sda4
## decreases size
of a logical volume
lvreduce -L
-
lvreduce -L
-
|
Filesystem
----------
commands:
|
Creating
|
## creates a
filesystem on top of lv
mkfs -t
## other way
mkfs.
-c --> checks for bad blocks before the
file system is created
-v --> verbose output
-l
-t
-m
to formats a device
into a specific filesystem:
## creates a swap
filesystem
mkswap
## creates an ext2
filesystem
mke2fs
-j
|
|
Resizing
|
## resizes a file
system
e2fsck -f
/dev/
resize2fs
/dev/
resize2fs -f
/dev/
resize_reiserfs
/dev/
|
|
Modifying
|
## adds label to a
storage device
e2label
tune2fs -L
## prints the label
of a storage device
e2label
|
Troubleshooting
---------------
|
Missing PV/VG/LV
|
One possible reason
if your LVM device doesn't show up in lvm commands like pvs, lvs, or vgs is
that the underlying LUN/storage device got vanished, corrupted, or its LUN
assignment has changed.
Example of the
error in /var/log/messages is:
Dec 24 19:56:42
orcl01 kernel: sd 2:0:8:0: Warning! Received an indication that the LUN
assignments on this target have changed. The Linux SCSI layer does not
automatically remap LUN assignments.
On possible
resolution is to scan the LVM device again
# Take backup of
your data.
# umount
/u01/LocalBackupIBM
# pvscan
# vgscan
# lvscan
# Run vgs & lvs
( If you can't see the volume group then you can proceed with the next step)
# vgcfgrestore -f
/etc/lvm/archive/orcl01_backupsIBM_00001-493843090.vg orcl01_backupsIBM
# lvs
|
|
Unable to remove LV
|
Issue:
You see this error
when remove LV via `lvremove`
Logical volume iscsi_vg/iscsi_lv is used by
another device.
Resolution:
1. Find out what is
holding the LV and take note of the major and minor numbers.
[root@node1 ~]#
dmsetup info -c
Name Maj Min Stat Open Targ
Event UUID
ha_vg-ha_lv 253
3 L--w 0 1
0 LVM-1C5GryiDYDe0lllqq1O8f9HsJyTpoUbe8OgaMOFiV5HEG4bi5fzzXLNHubN6A7pw
VolGroup00-LogVol01
253 1 L--w 2
1 0
LVM-SA8LTUF2yzFEV1RdgThw0ZiRxhyHFKuUIAjIC6ScnMtvH67BTyTN31hd82bgDSzd
VolGroup00-LogVol00
253 0 L--w 1
1 0
LVM-SA8LTUF2yzFEV1RdgThw0ZiRxhyHFKuUj6b8IVKEw37bTwOqy81Ud3juFCSJBg12
iscsi_vg-iscsi_lv 253
2 L--w 1 1
0 LVM-9ZTlpYoUyX9xLbFn2NoTfp3313n89MMh81CFS2y999Cvv9DqQbxdZ9XUnI3xx0QG
[root@node1 ~]#
2. Delete it
manually.
[root@node1 ~]# ls
-l /sys/dev/block/253\:2
lrwxrwxrwx. 1 root
root 0 Jun 15 06:37 /sys/dev/block/253:2 ->
../../devices/virtual/block/dm-2
[root@node1 ~]#
dmsetup remove /dev/dm-3
[root@node1 ~]#
3. You can now
remove the LV successfully.
[root@node1 ~]#
lvremove /dev/mapper/iscsi_vg-iscsi_lv -f
Logical volume "iscsi_lv"
successfully removed
[root@node1 ~]#
|
|
removing a stale
device
|
similar errors you
may encounter:
/dev/test_vg/test_lv:
read failed after 0 of 4096 at 5368643584: Input/output error
/dev/sdb: read
failed after 0 of 4096 at 0: Input/output error
cause:
physical device was
remove before deleting from OS
solution:
dmsetup remove
/dev/test_vg/*
echo 1 >
/sys/block/sdb/device/delete
or
reboot
|
Tutorials
---------
|
technique in
extending lvs by using physical extents
|
--> this will
prevent you from wasting space in your disk
sample extension
using "-L +50G" traditonal human readable size (you can see 1.34G
wasted or not used during lv extension)
[root@rt01 /]# pvs
PV
VG Fmt Attr PSize
PFree
/dev/sda2
VolGroup00 lvm2 a- 50.38G 0
/dev/sdb
VolGroup00 lvm2 a- 51.97G 1.34G
[root@rt01 /]#
extending using
physical extents:
1. determine first
if how many remaining free extents you have by extending the lv above the
maximum size
[root@rt01 /]#
lvextend -L +100g /dev/VolGroup00/lv_root --> here we extend it 100GB
which is above 50GB
Extending logical volume lv_root to 201.00
GB
Insufficient free space: 3200 extents
needed, but only 1642 available --> here you will see 1642 extents
available
[root@rt01 /]#
2. use the value in
#1 to extend your lv by using "-l" flag instead of "-L"
[root@rt01 /]#
lvextend -l +1642 /dev/VolGroup00/lv_root
Extending logical volume lv_root to 152.31
GB
Logical volume lv_root successfully resized
[root@rt01 /]#
3. check now your
remaining space (you can see no space is wasted)
[root@rt01 /]# pvs
PV
VG Fmt Attr PSize
PFree
/dev/sda2
VolGroup00 lvm2 a- 50.38G 0
/dev/sdb
VolGroup00 lvm2 a- 51.97G 0
[root@rt01 /]#
|
|
resizing a a single
vmdk disk
|
1. edit virtual
disk in vmware settings
2. reboot host
3. create new
partition
# fdisk /dev/sda
# n > primary > follow defaults
# t >
# w
4. reboot (at this
point I don't know if it is possible not to reboot)
5. pvcreate
/dev/sda4
6. vgextend vg_root
/dev/sda4
7. lvextend
/dev/mapper/vg_ftp01-lv_root /dev/sda4
8. resize2fs
/dev/mapper/vg_ftp01-lv_root
|
|
expanding a non-LVM
partition
|
1. shutdown server
2. add storage
3. insert gparted
cd into cdrom
4. boot server from
cd
5. use gparted to
expand partition
http://rbgeek.wordpress.com/2013/07/29/increase-the-root-filesystem-in-linux-inside-vmware-workstationvsphere-without-lvm/
## reminders in
designing your partitions
when layouting
partitions, make sure to make / the last partition
BAD layout:
/dev/sda1 = /
/dev/sda2 = /boot
/dev/sda3 = swap
--> in this
scenario, / is the first partition
--> there is
another partition right next to it
--> extending /
will be difficult because you must consider the starting and end boundaries
of the cylinder
GOOD layout:
/dev/sda1 = /boot
/dev/sda2 = swap
/dev/sda3 = /
--> there must
be no partitions after /
--> in that
case, you can extend it without worrying about the neighboring cylinder
|
|
Destroying an LVM
filesystem
|
1. Remove LV
lvremove
2. Deactivate and
remove VG
vgchange -a n
vgremove
3. Remove PV
pvremove
|
|
Resizing tmpfs
|
tmpfs is a RAM
based temporary file system which is generally mounted on /dev/shm. To to
extend the tmfs file system use below steps:
1. Check initial
size: # df -h /dev/shm
2. Edit fstab and
specify tmpfs size
# vi /etc/fstab
tmpfs /dev/shm tmpfs size=4g 0 0
3. Remount: #
mount -o remount /dev/shm
4. Check final
size: # df -h /dev/shm
|
|
Moving data to
another disk without any downtime
|
If you want to
migrate your data to a new disc - let's say you decided to
purchase another
storage appliance and you want your MYSQL datadir to move to
that appliance
without any downtime, LVM mirroring is the way to go.
1. Prepare your LUN
and assign/zone it to your server
2. Once the LUN is
available, add it to the current VG that holds the MYSQL
datadir
# vgextend myql_vg
/dev/sdc
3. Create a mirror
of the MYSQL logical volume pointing to the newly added LUN
(this is the RIGHT
side of the mirror)
# lvconvert -m1
/dev/mapper/mysql_vg-mysql_lv /dev/sdc
4. Wait until the
sync is complete (must be 100% under Cpy%Sync field)
# lvs
5. Once the sync is
complete, remove/detach the LEFT side of the mirror (that is
the old LUN) from
the MYSQL logical volume
# lvconvert -m0
/dev/mapper/mysql_vg-mysql_lv /dev/sdb
6. Remove the old
LUN from the MYSQL VG
# vgreduce mysql_vg
/dev/sdb
7. Remove PV label
# pvremove /dev/sdb
8. Detach it from
your server
9. Hola! You were
able to move your MYSQL database to a new and faster disk
without a downtime!
|
|
Renaming a volume
group in RHEL/CENTOS 6
|
1. Verify old name
# vgs
2. Rename VG
#
vgrename old_vg_name new_vg_name
3. Backup and
replace references in fstab
# cp
-p /etc/fstab /etc/fstab.$(date +%F)
#
grep old_vg_name /etc/fstab
/dev/mapper/old_vg_name-lv_root / ext4 defaults 1 1
/dev/mapper/old_vg_name-lv_swap swap swap defaults 0 0
4. Backup and
replace referencs in grub.conf
# cp
-p /boot/grub/grub.conf /boot/grub/grub.conf.$(date +%F)
#
grep old_vg_name /boot/grub/grub.conf
kernel /vmlinuz-2.6.32-642.1.1.el6.x86_64 ro
root=/dev/mapper/old_vg_name-lv_root ... rd_LVM_LV=old_vg_name/lv_root ...
rd_LVM_LV=old_vg_name/lv_swap ...
5. Backup and
rebuild initramfs
# cp
/boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r).img.bak
#
dracut -f
6. Restart server
#
reboot
|
No comments:
Post a Comment