March 15, 2015 at 11:10Correct increase in disk size in a virtual machine
Without pretending to be complete, I still think that this can be useful to system administrators.
The disk size increase in the virtual machine occurred with the following input: qcow2 virtual file format, the virtual machine uses lvm and ext4, the root partition is in the extended partition. The action usually occurs at night, when the load is minimal and downtime does not put much strain on the nerves. Although when working with highload projects, adrenaline still stands out enough to think 10 times before doing anything. Therefore, before starting the process, it is better to turn off the SMS notification system so as not to frighten colleagues with messages like “Server down” in the middle of the night.
1. Turn off the virtual machine
I did this via the GUI by clicking on the red power button in virt-manager. If there is no virt-manager, this can be done by issuing the shutdown command on the command line of the virtual machine.
2. On the hypervisor, increase the file size (in my case, by 200 gigabytes)
3. We cling the disk to another (service) virtual machine through the managing machine with virt-manager, an alternative is to boot from a CD with lvm support.
Accordingly, when booting from LiveCD, vdb changes to vda
4. We start the service machine (it should also have lvm on it) via virt-manager.
5. Next, on a service (or LVM liveCD) machine:
get the disk size:
increase the extended partition, if this is not done we get Error: Can't have overlapping partitions. ubuntu parted -gparted
increase the logical root partition
Now you need to increase the size of the physical disk in lvm
increase the size of the logical disk in lvm
The output of resize2fs should look like this:
Now check and fix the file system:
the disk is ready
6. turn off the service machine , disconnect the disk from it in virt-manager
from the command line, without using the GUI to manage virtual machines, you can do this with virsh, the use of which is well described here: managing virtual machines from the command line
7. Start the server
An increase with a minimum downtime, almost on the fly, checked on lvm2 / ext4 can be done as follows:
1. An increase in file size by 200 gigabytes is performed on the hypervisor
2. Rebooting the virtual machine
3. On the virtual machine
Let's see the size of the physical disk and all logical partitions
increase extended partition
increase the logical root partition
Now you need to increase the size of the physical disk in lvm
increase the size of the logical disk in lvm
In this case, you cannot check and fix the file system, fsck -f / dev / mapper / vm - db - 0 - vg-root will kill the file system
Check what happened:
Continuation of the topic
The disk size increase in the virtual machine occurred with the following input: qcow2 virtual file format, the virtual machine uses lvm and ext4, the root partition is in the extended partition. The action usually occurs at night, when the load is minimal and downtime does not put much strain on the nerves. Although when working with highload projects, adrenaline still stands out enough to think 10 times before doing anything. Therefore, before starting the process, it is better to turn off the SMS notification system so as not to frighten colleagues with messages like “Server down” in the middle of the night.
1. Turn off the virtual machine
I did this via the GUI by clicking on the red power button in virt-manager. If there is no virt-manager, this can be done by issuing the shutdown command on the command line of the virtual machine.
2. On the hypervisor, increase the file size (in my case, by 200 gigabytes)
qemu-img resize /path/to/vm-disk.qcow2 +200G3. We cling the disk to another (service) virtual machine through the managing machine with virt-manager, an alternative is to boot from a CD with lvm support.
Accordingly, when booting from LiveCD, vdb changes to vda
4. We start the service machine (it should also have lvm on it) via virt-manager.
5. Next, on a service (or LVM liveCD) machine:
parted /dev/vdbget the disk size:
(parted) print
Model: Virtio Block Device (virtblk)
Disk /dev/vdb: 1288GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number Start End Size Type File system Flags
1 1049kB 256MB 255MB primary ext2 boot
2 257MB 1000GB 1000GB extended
5 257MB 1000GB 1000GB logical lvmincrease the extended partition, if this is not done we get Error: Can't have overlapping partitions. ubuntu parted -gparted
(parted) resizepart 2
End? [1000GB]? 1288Gbincrease the logical root partition
(parted) resizepart 5
End? [1000GB]? 1288Gb
(parted) qNow you need to increase the size of the physical disk in lvm
pvresize /dev/vdb5increase the size of the logical disk in lvm
root@vm-service:/etc# lvextend /dev/vm-db-0-vg/root -l +100%FREE
lvextend /dev/vm-db-0-vg/root -l +100%FREE
File descriptor 7 (pipe:[7918]) leaked on lvextend invocation. Parent PID 1378: bash (на это можно не обращать внимания)
Extending logical volume root to 1.12 TiB
Logical volume root successfully resized
root@vm-service:/etc# resize2fs /dev/vm-db-0-vg/root
The output of resize2fs should look like this:
The filesystem on /dev/vm-db-0-vg/root is now 231278592 blocks long.Now check and fix the file system:
fsck -f /dev/mapper/vm--db--0--vg-root
the disk is ready
6. turn off the service machine , disconnect the disk from it in virt-manager
from the command line, without using the GUI to manage virtual machines, you can do this with virsh, the use of which is well described here: managing virtual machines from the command line
7. Start the server
An increase with a minimum downtime, almost on the fly, checked on lvm2 / ext4 can be done as follows:
1. An increase in file size by 200 gigabytes is performed on the hypervisor
qemu-img resize /path/to/vm-disk.qcow2 +200G2. Rebooting the virtual machine
3. On the virtual machine
parted /dev/vdaLet's see the size of the physical disk and all logical partitions
(parted) print
Model: Virtio Block Device (virtblk)
Disk /dev/vda: 1288GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number Start End Size Type File system Flags
1 1049kB 256MB 255MB primary ext2 boot
2 257MB 1000GB 1000GB extended
5 257MB 1000GB 1000GB logical lvmincrease extended partition
(parted) resizepart 2
End? [1000GB]? 1288Gbincrease the logical root partition
(parted) resizepart 5
End? [1000GB]? 1288Gb
(parted) qNow you need to increase the size of the physical disk in lvm
pvresize /dev/vda5increase the size of the logical disk in lvm
root@vm-db-0:/etc# lvextend /dev/vm-db-0-vg/root -l +100%FREE
lvextend /dev/vm-db-0-vg/root -l +100%FREE
File descriptor 7 (pipe:[7918]) leaked on lvextend invocation. Parent PID 1378: bash
Extending logical volume root to 1.12 TiB
Logical volume root successfully resized
root@vm-db-0:/etc# resize2fs /dev/vm-db-0-vg/rootIn this case, you cannot check and fix the file system, fsck -f / dev / mapper / vm - db - 0 - vg-root will kill the file system
Check what happened:
df -hContinuation of the topic