Skip to content

Latest commit

 

History

History
196 lines (141 loc) · 13.9 KB

G906 - Appendix 06 ~ Handling VM or VM template volumes.md

File metadata and controls

196 lines (141 loc) · 13.9 KB
Raw

G906 - Appendix 06 ~ Handling VM or VM template volumes

The hard disks attached to a VM or a VM template are, in fact, VM disk images. In the case of the Debian VM created in the G020 guide, its hard disk is an image that has been created as a LVM light volume within a thinpool. This way, the disk image is not just a file, but a virtual storage device that contains the VM's entire filesystem. How to locate and, when necessary, handle such an image? Read the following subsections to have a glimpse about this.

Installing the libguestfs-tools package

The Proxmox VE web console only gives you a very limited range of actions, like creation or size enlargement, to perform over any VM's hard disks. Also, there's in your system the qemu-img command to manipulate these images, but it's also kind of limited. A much more powerful command toolkit for handling VM disk images is the one provided by the libguestfs-tools package.

  1. You don't have it installed in your Proxmox VE host, so open a shell in it and install the package with apt.

    $ sudo apt install -y libguestfs-tools

    This package's installation will execute a considerable number of actions and install several dependencies, so you'll see a lot of output lines on this process.

  2. Since this installation has done quite a bunch of things in your Proxmox VE host, its better if you reboot it right after installing the libguestfs-tools package.

    $ sudo reboot

The libguestfs-tools package comes with a big set of commands that allows you to handle in complex ways VM disk images. Check them out in the documentation available at libguestfs' official page.

Locating and checking a VM or VM template's hard disk volume

So, where in your system is the VM disk image of your VM template?

  1. In the Proxmox VE web console, go to your VM template's Hardware view, and read the Hard Disk line.

    VM template's Hardware view

    Remember the ssd_disks:base-100-disk-0 string: it's the name of the hard disk volume within your Proxmox VE node.

  2. Next, open a shell terminal (as your administrator user) on your Proxmox VE host and execute the following lvs command.

    $ sudo lvs -o lv_full_name,pool_lv,lv_attr,lv_size,lv_path,lv_dm_path
  LV                       Pool      Attr       LSize   Path                          DMPath
  hddint/hdd_data                    twi-a-tz-- 870.00g                               /dev/mapper/hddint-hdd_data
  hddint/hdd_templates               -wi-ao----  60.00g /dev/hddint/hdd_templates     /dev/mapper/hddint-hdd_templates
  hddusb/hddusb_bkpdata              twi-a-tz--  <1.31t                               /dev/mapper/hddusb-hddusb_bkpdata
  hddusb/hddusb_bkpvzdumps           -wi-ao---- 520.00g /dev/hddusb/hddusb_bkpvzdumps /dev/mapper/hddusb-hddusb_bkpvzdumps
  pve/root                           -wi-ao---- <37.50g /dev/pve/root                 /dev/mapper/pve-root
  pve/swap                           -wi-ao----  12.00g /dev/pve/swap                 /dev/mapper/pve-swap
  ssdint/base-100-disk-0   ssd_disks Vri---tz-k  10.00g /dev/ssdint/base-100-disk-0   /dev/mapper/ssdint-base--100--disk--0
  ssdint/ssd_disks                   twi-aotz-- 880.00g                               /dev/mapper/ssdint-ssd_disks

    In the lvs output above, you can see your VM template's hard disk volume named as ssdint/base-100-disk-0. This means that it's a volume within the ssdint LVM volume group you created back in the G005 guide. Not only that, in the Pool column you see the name ssd_disks, which refers to the LVM thinpool you created in the G019 guide. The next Attr column gives you some information about the volume itself:

    • V indicates that the volume is virtual.
    • r means that this volume is read-only.
    • i refers to the storage allocation policy used in this volume, in this case is inherited.
    • t means that this volume uses the thin provisioning driver as kernel target.
    • z indicates that newly-allocated data blocks are overwritten with blocks of (z)eroes before use.
    • k is a flag to make the system skip this volume during activation.

    NOTE
    To know all the possible values in the Attr column, check the Notes section of the lvs manual (command man lvs).

    Next to the Attr column, you can see the size assigned to the volume, 10 GiB in this case, although this number is just logical. If you return to the Proxmox VE web console, and then go to the ssd_disks thinpool Summary view, you'll see in Usage that much less than 10 GiB are actually in use.

    Real usage of ssd_disks thinpool

    In this case, just 1.98 GiB are actually in use in the thinpool and, at this point, only the base-100-disk-0 volume is present there.

    And what about the columns Path and the DMPath of the lvs output? They're the paths to the handler files used by the system to manage the light volumes. You can see them with the ls command, except the ones used for the base-100-disk-0 volume. Since this volume is not active (remember the k flag in the Attr column), you won't find the corresponding files present in the system.

    In conclusion, with the storage structure you have setup in your system, mostly based on LVM thinpools, all your hard disk volumes will be virtual volumes within LVM thinpools. In the case of your VM template's sole hard disk, the concrete LVM location is as follows:

    • Volume Group ssdint.
    • Thinpool ssd_disks.
    • Light Volume base-100-disk-0.

    Remember that, in this scenario, the ssdint VG corresponds to an entire physical LVM volume set within the /dev/sda4 PV which, moreover, shares the real underlying ssd unit with the /dev/sda3 PV (the one containing the pve volume group for the Proxmox VE system volumes).

You've seen the LVM side of the story, but you can get more information about the VM disk image by using some libguestfs commands. In fact, you can even get inside the filesystem within the disk image. To do so, first you have to activate the disk image as light volume, since when you turned the VM into a template, it's hard disk is now disabled and read-only for LVM to avoid further modifications.

  1. Reactivate the volume with the following lvchange command.

    sudo lvchange -ay -K ssdint/base-100-disk-0

    Since the previous command doesn't give back any output in success, use the lvs command to verify it's status.

    $ sudo lvs ssdint/base-100-disk-0
  LV              VG     Attr       LSize  Pool      Origin Data%  Meta%  Move Log Cpy%Sync Convert
  base-100-disk-0 ssdint Vri-a-tz-k 10.00g ssd_disks        18.56

    Notice the a among the values under the Attr column, that means the volume is now active. Other command to check out what light volumes are active is lvscan.

    $ sudo lvscan
  ACTIVE            '/dev/hddusb/hddusb_bkpvzdumps' [520.00 GiB] inherit
  ACTIVE            '/dev/hddusb/hddusb_bkpdata' [<1.31 TiB] inherit
  ACTIVE            '/dev/hddint/hdd_templates' [60.00 GiB] inherit
  ACTIVE            '/dev/hddint/hdd_data' [870.00 GiB] inherit
  ACTIVE            '/dev/ssdint/ssd_disks' [880.00 GiB] inherit
  ACTIVE            '/dev/ssdint/base-100-disk-0' [10.00 GiB] inherit
  ACTIVE            '/dev/pve/swap' [12.00 GiB] inherit
  ACTIVE            '/dev/pve/root' [<37.50 GiB] inherit

    This command shows you the state of all the present light volumes, active or inactive. Also notice that it shows the full handler path to the volumes.

  2. With the base-100-disk-0 volume now active, you can check out the status of the filesystem it contains with the libguestfs command virt-df.

    $ sudo virt-df -h -a /dev/ssdint/base-100-disk-0
Filesystem                                Size       Used  Available  Use%
base-100-disk-0:/dev/sda1                 469M        47M       398M   11%
base-100-disk-0:/dev/debiantpl-vg/root
                                          8.3G       1.1G       6.8G   14%

    This virt-df command is very similar to the df, and gives you a view of how the storage space is used in the filesystem contained in the volume. Remember that this is the filesystem of your Debian VM template, and it also has a swap volume although the virt-df command doesn't show it.

  3. Another command you might like to try and get a much more complete picture of the filesystem within the base-100-disk-0 volume is virt-filesystems.

    $ sudo virt-filesystems -a /dev/ssdint/base-100-disk-0 --all --long --uuid -h
Name                     Type       VFS  Label MBR Size Parent            UUID
/dev/sda1                filesystem ext2 -     -   469M -                 ec1735f0-edf0-41c5-b54f-9012092a7a2c
/dev/debiantpl-vg/root   filesystem ext4 -     -   8.3G -                 41032d28-f6d8-416b-936b-bb0fd803e832
/dev/debiantpl-vg/swap_1 filesystem swap -     -   976M -                 71cb513a-ef5f-480c-ad54-dea7734d9a97
/dev/debiantpl-vg/root   lv         -    -     -   8.5G /dev/debiantpl-vg 0NOJ1z-HMPD-nHl6-OdQr-rgpR-nxtf-FI0i44
/dev/debiantpl-vg/swap_1 lv         -    -     -   976M /dev/debiantpl-vg Kci3BQ-ITxT-SNuC-ITux-BI1x-Wtr2-t9c2RG
/dev/debiantpl-vg        vg         -    -     -   9.5G /dev/sda5         8cH6WfvK16w50hvYRVx1PyuB2gQrxLgW
/dev/sda5                pv         -    -     -   9.5G -                 Cx7CAvy5XaYDMWtdVERRC4f7aTfG2jDQ
/dev/sda1                partition  -    -     83  487M /dev/sda          -
/dev/sda2                partition  -    -     05  1.0K /dev/sda          -
/dev/sda5                partition  -    -     8e  9.5G /dev/sda          -
/dev/sda                 device     -    -     -   10G  -                 -

    See how not only this command returns info about the two LVM light volumes (root and swap_1), but also shows the sda partitions and the LVM physical volume sda5.

  4. After you've finished checking the base-100-disk-0 volume, it's better if you deactivate it. For this, use lvchange again.

    $ sudo lvchange -an ssdint/base-100-disk-0

    Verify its inactive status directly with the lvscan command.

    $ sudo lvscan
  ACTIVE            '/dev/hddusb/hddusb_bkpvzdumps' [520.00 GiB] inherit
  ACTIVE            '/dev/hddusb/hddusb_bkpdata' [<1.31 TiB] inherit
  ACTIVE            '/dev/hddint/hdd_templates' [60.00 GiB] inherit
  ACTIVE            '/dev/hddint/hdd_data' [870.00 GiB] inherit
  ACTIVE            '/dev/ssdint/ssd_disks' [880.00 GiB] inherit
  inactive          '/dev/ssdint/base-100-disk-0' [10.00 GiB] inherit
  ACTIVE            '/dev/pve/swap' [12.00 GiB] inherit
  ACTIVE            '/dev/pve/root' [<37.50 GiB] inherit

    Remember that when the volume is inactive, the handler file doesn't exist in the system. Also, know that there is a corresponding /dev/mapper handler file for each volume. For your base-100-disk-0 volume, the full path would be /dev/mapper/ssdint-base--100--disk--0.

Relevant system paths

Directories on Proxmox VE host

  • /dev
  • /dev/mapper
  • /dev/ssdint

Files on Proxmox VE host

  • /dev/mapper/ssdint-base--100--disk--0
  • /dev/ssdint/base-100-disk-0

References

Handling of VM disk images

Navigation

<< Previous (G905. Appendix 05) | +Table Of Contents+ | Next (G907. Appendix 07) >>