How can I increase the size of the root partition of a system at runtime?

I have a partition that is not allocated after the root partition (which is also ext4), how can I add that unallocated space to the space allocated to the root partition without having to shutdown the server?

  • 3
    I hate to sound like a stick in the mud, but this entails a fair bit of risk? Why does this need to happen? Is uptime the main constraint? – DevNull Sep 30 '13 at 17:28
  • You can't resize a partition to the left, because that would actually be a move. – Zaz May 17 '15 at 15:16
  • 2
    Increasing the size of ext4 parititions online is easy. The difficult part would be shrinking (your headline is about "resizing"). For people interested by ANY manipulation on a root partition (move, shrink, change filesystem, device) at runtime should consult my answer: askubuntu.com/a/728141/21888 – vaab Feb 1 '16 at 2:28

10 Answers 10

GUI (Ubuntu 14.04 and later): GParted v0.17 and later provide a nice GUI for this. (Older versions will refuse to resize a mounted partition).

Command line (any Ubuntu version): There are three steps to this.

Step 1. The partition must first be resized. If you're using LVM, it's easy, and you presumably know how to proceed. If you're using classic partitions, it's a bit more complicated, and may require a reboot (though you never have to boot another system or live CD).

This is how I do it: Use fdisk to first delete the partition (the idea is that the data on disk will be preserved), then carefully recreate it with a larger size at the same position.

Example:

$ sudo fdisk /dev/sda

Command (m for help): p

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *        2048     9437183     4717568   83  Linux

Command (m for help): d
Selected partition 1

Command (m for help): p

   Device Boot      Start         End      Blocks   Id  System

Command (m for help): n
Command action
   e   extended
   p   primary partition (1-4)
p
Partition number (1-4, default 1): 1
First sector (2048-10485759, default 2048):
Using default value 2048
Last sector, +sectors or +size{K,M,G} (2048-10485759, default 10485759):
Using default value 10485759

Command (m for help): p

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1            2048    10485759     5241856   83  Linux

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 or after you run partprobe(8) or kpartx(8)
Syncing disks.

Again, it is critical that the new partition starts at the same block as the old. The Id should also match (83 for Linux systems). Be prepared to lose all your data at the slightest typo.

To be on the safe side, you may also restore the boot flag (which according to Wikipedia is still required on some computers) by pressing a.

See the comment section for what to do if your swap partition is in the way.

By now it should be apparent why people recommend using a live CD. ;-)

Step 2. As fdisk helpfully reminds you, you must reload the partition table before proceeding. The safest way is to simply reboot; but you can also use partprobe or kpartx (more information).

Step 3. Once the partition is resized and the partition table reloaded, it's a simple matter of running resize2fs on the file system, and you can do this even when it's mounted as the root partition.

Example:

$ sudo resize2fs /dev/sda1
  • 13
    This worked perfectly for me. However I did additionally ensure that the boot flag kept is original state. – Augustus Kling Aug 10 '12 at 20:56
  • 4
    @jbo5112: As fdisk says, partprobe or kpartx may work instead of a reboot; see also this question. Even if you reboot, the solution is still preferable to using a live CD when it comes to downtime, where a simple reboot can be less than 10 s for a virtual machine. It's also faster in operator time, which is why I usually use this approach myself. :) – Søren Løvborg Dec 3 '13 at 18:07
  • 1
    @Raymond: If memory pressure allows (see free -h), disable the swap (swapoff /dev/sda2), change the partition table (including deleting and recreating the swap partition) and either 1) reboot or 2) reload the partition table and swapon again. (If memory's too tight to disable swap temporarily, you can still create and enable a new swap partition (/dev/sda3), then swapoff sda2; but then you'll have to update /etc/fstab with the new swap device name.) – Søren Løvborg Jul 8 '15 at 21:15
  • 1
    If you are using vmware and have extended the size of the disk, you will have to run sudo lshw -C disk to rescan the filesystems so the vm recognises the bigger drive. Then follow the instructions above. – Guy Apr 26 '16 at 8:23
  • 1
    What about shrinking? – Aaron Franke Jul 28 at 7:09

It is possible to do a on-line resize of a ext4 filesystem, even if it's your root partition. Use the resize2fs command.

sudo resize2fs /dev/sda1

EDIT: On-line shrinking is not allowed:

root@brunojcm-htpc:/home# resize2fs /dev/sda5 2654693
resize2fs 1.42 (29-Nov-2011)
Filesystem at /dev/sda5 is mounted on /; on-line resizing required
resize2fs: On-line shrinking not supported
  • 50
    From man resize2fs: The resize2fs program does not manipulate the size of partitions. If you wish to enlarge a filesystem, you must make sure you can expand the size of the underlying partition first. This can be done using fdisk(8) by deleting the partition and recreating it with a larger size or using lvextend(8),if you're using the logical volume manager lvm(8). This question is about resizing the partition, not the filesystem. The distinction is subtle but very important. – Eliah Kagan Jun 3 '12 at 7:07
  • 8
    You can use fdisk to delete the root partion and then recreate it at the same starting block. fdisk will write out the change, but it won't take effect till after a reboot. after the reboot you can use the resize2fs program to send the disk to fill the partion. – James Becwar Jun 14 '12 at 15:15
  • 3
    I have just resized an ext4 root partition online. Therefore I can confirm it's possible. But instead of passing /dev/sda* as parameter to resize2fs, you need to pass the logical volume name. – CDR Dec 4 '12 at 22:37
  • 8
    I find the first paragraph of the resize2fs manpage most interesting for the initial question: The resize2fs program will resize ext2, ext3, or ext4 file systems. It can be used to enlarge or shrink an unmounted file system located on device. If the filesystem is mounted, it can be used to expand the size of the mounted filesystem, assuming the kernel supports on-line resizing. (As of this writing, the Linux 2.6 kernel supports on-line resize for filesystems mounted using ext3 and ext4.). – mo' Dec 23 '12 at 14:33
  • 3
    Please don't muck with fdisk when growpart will do this very easily for you. – STRML Sep 16 '17 at 13:45

An easier solution - use growpart <device> <partition>:

growpart /dev/xvda 1  # Grows the partition; note the space
resize2fs /dev/xvda1  # Grows the filesystem

As always, back up your partition table (sfdisk -d /dev/xvda > partition_bak.dmp) just in case.

  • What about shrinking? – Aaron Franke Jul 28 at 7:09
  • Was hoping to use this method, but got sfdisk: /dev/nvme0n1p5: does not contain a recognized partition table (using dual boot with windows) – monkut Jul 28 at 23:26
  • Shrinking is not possible to do on-line. I recommend using gparted for it. – STRML Jul 29 at 15:42
  • This was perfect for resizing the root partition and filesystem of my AWS VM. Cheers. – MetaFight Aug 10 at 0:00
  • growpart is part of cloud-utils. In case you don't have installed, you can install with apt-get install cloud-utils – klor Aug 21 at 19:50

You could also just make use of GParted - as long as the partition you are resizing is not the one you booted from - else I suggest the live CD option is somewhat easier for newbies.

GParted basically does all of the steps - just based on a GUI fronted.

  • I booted my system with Ubuntu 12.04 Live CD and i resized ext4 partition with GParted. Worked good for me. Anyway, before this operation I backed up all my important data. – StandDuPp May 28 '13 at 20:16
  • i think, gparted needs partition unmounted. but i can be wrong. – Nick Jan 27 '16 at 18:02
  • The question is obviously about the partition they booted from, and booting a live CD requires restarting the machine. -1 – wjandrea May 21 '17 at 2:28

Yes, you can shrink/move/grow an online root partition without any reboots (nor livecd, nor usbkey): consult this answer. It's very well written and easy to follow, although quite long and a little risky. So if you only want to grow your ext4 partition, you can stick to the conventional working resize2fs solutions.

The general solution I've lnked will work on any type of dedicated or VPS solution for instance.

TLDR; this solution implies to pivot_root to tmpfs so you can umount safely your root partition live and fiddle with it. Once done, you'll pivot_root back on your new root partition.

This allows pretty much any manipulation on the root file system (move it, change filesystem, changing it's physical device...).

No reboot are required in the process, and this allows to bypass limitation of resize2fs not being able to shrink ext4 partitions.

I have personally used this, and it works very well on debian system also, so it should work on Ubuntu. I'm very surprised not to see this in-depth solution a little more linked to the many question in stackexchange web sites that deals with the same issue.

Note: Of course if you want to grow your partition, a simple resize2fs will be enough as stated in numerous places and in other answers here.

  • 4
    I think to most people, once you've stopped all programs and services accessing the root partition, you might as well have rebooted the machine. For shrinking/moving, that's may still be faster that using a live CD, but for growing (by far the most common task, and what OP asked about), there are ways that don't involve a temporary shutdown of most of the system. – Søren Løvborg Mar 8 '16 at 18:35
  • 2
    @SørenLøvborg: You can restart the core services that needs continuous production while doing the full procedure. There are many configurations where you can't put LiveCD (VPS instances, dedicated servers... ) or circumstances where you want to avoid any reboots for specific reasons. The original question's title mentions "resizing", which attracts people looking for shrinking partitions online. **No other solution allows shrinking ext4 online.**This solution is risky, complex, but the most powerful of all and it fills the shortcomings of the others. – vaab Mar 9 '16 at 5:41
  • Please do not post answers depending on external links. Put the relevant part into your answer or post the link as a comment to the question. See How to Answer for details. – Melebius Oct 12 at 11:44

Just finished resizing an ext4 root partition on a live system while the root was mount.

[root@habib i686]# resize2fs /dev/vg_habib/lv_root
resize2fs 1.42 (29-Nov-2011)
Filesystem at /dev/vg_habib/lv_root is mounted on /; on-line resizing required
old_desc_blocks = 4, new_desc_blocks = 10
Performing an on-line resize of /dev/vg_habib/lv_root to 38427648 (4k) blocks.
The filesystem on /dev/vg_habib/lv_root is now 38427648 blocks long.

[root@habib i686]# 

I just did this successfully without umount, pivot_root, or temporary removal of the main partition, using parted 3.2 on Ubuntu 16.04, 4.4.0 kernel. To be cautious I did everything from a virtual console with networking disabled, and took a snapshot beforehand just in case, but the snapshot wasn't needed, so I could have just as well done this via SSH and without changing runlevels.

Determine partition size: parted /dev/sda1 print | egrep "Disk.*GB"

Optionally switch to multi-user mode without networking (must be done from a console, not SSH):

runlevel     # remember the original runlevel
init 2

Optionally take a VM snapshot to be cautious.

Resize partition:

parted
p
resizepart NUMBER SIZE
# answer "Yes" when asked about resizing a live partition.
q

Resize filesystem: resize2fs /dev/sda1

If anything goes wrong, you could restore your snapshot here. If all went fine, return to normal runlevel (obtained above) - normally 5: init 5. It may be better to do a full reboot at this point to make sure everything comes back properly (I had a date/ntp issue afterward).

As stated before:

  • expanding live from a root system is possible.(no difficulties, as the boot section ain't to be moved)

  • shrinking a live root partition needs to be done from external boot device (boot from live system cd/usb-stick), as if there’s any fault, mismatch ..whatever..your system hangs, needs to be rebooted and will eventually not be able to boot correctly.

Any sort of "but I did it and it works" is pure luck.

Follow these steps.

  1. open terminal as superuser su
  2. run parted
  3. type p to see the available partitions
  4. identify your root partition number (ex: 'sda 3' means number 3) and delete an adjacent partition by using rm PARTITION NUMBERto create free space.
  5. now increase the root size by typing resizepart ROOT NUMBER and reboot system if needed
  6. exit parted by typing exit and in terminal type partprobe and hit enter (this can be done even after rebooting)
  7. finally run resize2fs /dev/sda PARTITION NUMBER and enjoy spacious root partition.

I would like to make an extension on the answer of @Søren Løvborg: extending the partition with a swap partition present.

First the layout of the disk after extending it:

$sudo parted /dev/sda 'unit s print' free
Model: ATA VBOX HARDDISK (scsi)
Disk /dev/sda: 14336000s
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:

Number  Start      End        Size       Type      File system     Flags
        63s        2047s      1985s                Free Space
 1      2048s      10485759s  10483712s  primary   ext4            boot
        10485760s  10487805s  2046s                Free Space
 2      10487806s  12580863s  2093058s   extended
 5      10487808s  12580863s  2093056s   logical   linux-swap(v1)
        12580864s  14335999s  1755136s             Free Space

So sda1 needs to be extended with the free space at the end of the disk, but the swap partition is in between them. This is how you can do it:

First we need to disable swap. Check how much it is used and if you can turn it off.

$ free -h
              total        used        free      shared  buff/cache   available
Mem:           992M         52M        464M        3.2M        475M        784M
Swap:          1.0G          0B        1.0G

swap is unused here so we can turn it off

$sudo swapoff /dev/sda2

Now we will change the partition table:

$sudo fdisk /dev/sda

(note: if you happen to have the first partition start at sector 63 instead of 2048, you need to add the option -c=dos)

Welcome to fdisk (util-linux 2.27.1).
Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.


Command (m for help): p
Disk /dev/sda: 6.9 GiB, 7340032000 bytes, 14336000 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x9e11c6df

Device     Boot    Start      End  Sectors  Size Id Type
/dev/sda1  *        2048 10485759 10483712    5G 83 Linux
/dev/sda2       10487806 12580863  2093058 1022M  5 Extended
/dev/sda5       10487808 12580863  2093056 1022M 82 Linux swap / Solaris

Command (m for help): d
Partition number (1,2,5, default 5): 2

Partition 2 has been deleted.

Command (m for help): d
Selected partition 1
Partition 1 has been deleted.

Command (m for help): n
Partition type
   p   primary (0 primary, 0 extended, 4 free)
   e   extended (container for logical partitions)
Select (default p): p
Partition number (1-4, default 1): 1
First sector (2048-14335999, default 2048):
Last sector, +sectors or +size{K,M,G,T,P} (2048-14335999, default 14335999): 122                                                                  42941

Created a new partition 1 of type 'Linux' and of size 5.9 GiB.

Command (m for help): n
Partition type
   p   primary (1 primary, 0 extended, 3 free)
   e   extended (container for logical partitions)
Select (default p): p
Partition number (2-4, default 2): 2
First sector (12242942-14335999, default 12242944):
Last sector, +sectors or +size{K,M,G,T,P} (12242944-14335999, default 14335999):

Created a new partition 2 of type 'Linux' and of size 1022 MiB.

Command (m for help): a
Partition number (1,2, default 2): 1

The bootable flag on partition 1 is enabled now.

Command (m for help): w
The partition table has been altered.
Calling ioctl() to re-read partition table.
Re-reading the partition table failed.: Device or resource busy

The kernel still uses the old table. The new table will be used at the next reboot or after you run partprobe(8) or kpartx(8).

Note: the size of sda1 is the total amount of sectors minus the sector size of your swap partition: 14335999-2093058=12242941

As fdisk mentions: the kernel is still using the old partition table so we need to reload it.

$partprobe

Now we need to run resize2fs on sda1 (do not forget this!)

$resize2fs /dev/sda1
resize2fs 1.42.12 (29-Aug-2014)
Filesystem at /dev/sda1 is mounted on /; on-line resizing required
old_desc_blocks = 3, new_desc_blocks = 10
The filesystem on /dev/sda1 is now 38833617 (4k) blocks long.

Now, things are not over yet. As you've probably noticed sda2 is partitioned as type Linux (Ext4). For some reason, there is no way in fdisk to choose the type. So we have to alternate it using cfdisk

$ sudo cfdisk

Choose sda2 and change type to 82 Linux swap / Solarisand make sure you write it (type yes to confirm)

Now we can re-activate the swap

$mkswap /dev/sda2
/dev/sda2
UUID=d58bf1cb-d27a-487d-b337-056767fd5ad6 none swap sw 0 0

And finally turn it on:

$swapon /dev/sda2

The only thing we need to do is to update fstab to mount the swap partition automatically upon booting

$sudo nano /etc/fstab

And change the UUID of the swap partition to the output above:

# swap was on /dev/sda5 during installation
UUID=d58bf1cb-d27a-487d-b337-056767fd5ad6 none            swap    sw              0       0

Now all is well and you can reboot without problems.

protected by David Foerster Sep 15 '17 at 9:13

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.