Manually Changing a Xen ECS to a KVM ECS (Linux)

Scenarios

Before changing a Xen ECS that runs Linux to a KVM ECS, install and configure required drivers.

This section describes how to manually install drivers on a Linux ECS, configure automatic disk attachment, and change Xen to KVM.

For instructions about how to use a script to automatically install drivers, see Automatically Changing a Xen ECS to a KVM ECS (Linux).

• Xen ECSs include S1, C1, C2, and M1 ECSs.
• To obtain KVM ECSs, see the Virtualization column in ECS Specifications.
• To support both Xen and KVM, Linux ECSs require the xen-pv and virtio drivers. Before changing a Xen ECS to a KVM ECS, make sure that the Linux ECS has been configured, including driver installation and automatic disk attachment.

Constraints

• The ECS needs to be stopped during the the specification modification, so you are advised to perform this operation during off-peak hours.
• To prevent data loss, the specifications of Linux ECSs that use LVM or RAID arrays cannot be modified.
• A Xen ECS with more than 24 VBD disks attached cannot be changed to a KVM ECS.
• A Xen ECS can be changed to a KVM ECS, but a KVM ECS cannot be changed to a Xen ECS.

Procedure

Figure 1 shows the flowchart for manually changing a Xen ECS to a KVM ECS.

Figure 1 Flowchart for manually changing a Xen ECS to a KVM ECS

Step 1: Create a System Disk Snapshot

If you modify the specifications of an ECS without installing the driver, the ECS may become unavailable and the data on the system disk may be lost. Therefore, create a snapshot for the system disk first.

1. Before you create a system disk snapshot, check the ECS.

   Stop and then start the ECS to ensure that services can run properly after the ECS is started.

2. For details about how to create a system disk snapshot, see Creating a Snapshot in Elastic Volume Service User Guide.

After the specifications are modified, manually delete the snapshot on the snapshot page if you have verified that services are running properly.

Step 2: Install and Configure Drivers

Perform the following operations to manually install drivers on an ECS.

1. Log in to the ECS.
2. Uninstall tools from the ECS.

   For details, see Uninstalling PV Drivers from a Linux ECS.

3. Change the GRUB disk ID to UUID.

   For details, see Changing the Disk Identifier in the GRUB Configuration File to UUID.

4. Change the fstab disk ID to UUID.

   For details, see Changing the Disk Identifier in the fstab File to UUID.

5. Install native Xen and KVM drivers.

   For details, see Installing Native Xen and KVM Drivers.

Step 3: Verify that the Drivers Have Been Installed

Perform the following operations to check whether the drivers have been installed and the configuration files have been modified.

Before manually modifying specifications, make sure that the ECS has been configured correctly.

1. Log in to the ECS.
2. Run the following command to check whether the root partition is in UUID format:

   cat /boot/grub/grub.cfg

   • If yes, the disk ID in the GRUB configuration file has been changed to UUID.
   • If no, the modification failed. In such a case, change the GRUB disk ID to UUID again by referring to Step 2: Install and Configure Drivers.

   ……menuentry 'Ubuntu Linux, with Linux 3.13.0-24-generic' --class ubuntu --class gnu-linux --class gnu --class os --unrestricted $menuentry_id_option 'gnulinux-3.13.0-24-generic-advanced-ec51d860-34bf-4374-ad46-a0c3e337fd34' {
   recordfail
   load_video
   gfxmode $linux_gfx_mode
   insmod gzio
   insmod part_msdos
   insmod ext2
   if [ x$feature_platform_search_hint = xy ]; then
   search --no-floppy --fs-uuid --set=root ec51d860-34bf-4374-ad46-a0c3e337fd34
   else
   search --no-floppy --fs-uuid --set=root ec51d860-34bf-4374-ad46-a0c3e337fd34
   fi
   echo 'Loading Linux 3.13.0-24-generic ...'
   linux /boot/vmlinuz-3.13.0-24-generic root=UUID=ec51d860-34bf-4374-ad46-a0c3e337fd34 ro
   echo 'Loading initial ramdisk ...'
   initrd /boot/initrd.img-3.13.0-24-generic
   }

   

   The path in which the GRUB configuration file is stored varies depending on the OS. For example, the path can be /boot/grub/menu.lst, /boot/grub/grub.cfg, /boot/grub2/grub.cfg, or /boot/grub/grub.conf.

3. Run the following command to check whether the disk ID in the fstab configuration file is UUID:

   cat /etc/fstab

   • If yes, the disk ID has been changed to UUID.
   • If no, the modification failed. In such a case, change the fstab disk ID to UUID again by referring to Step 2: Install and Configure Drivers.

   [root@****** ~]# cat /etc/fstab 
   UUID=4eb40294-4c6f-4384-bbb6-b8795bbb1130  /       xfs     defaults    0 0
   UUID=2de37c6b-2648-43b4-a4f5-40162154e135  swap    swap    defaults    0 0

4. Check whether the native Xen and KVM drivers have been installed.
   • If the boot virtual file system is initramfs, run the following commands:

     lsinitrd /boot/initramfs-`uname -r`.img | grep ` uname -r ` | grep xen

     lsinitrd /boot/initramfs-`uname -r`.img | grep ` uname -r ` |grep virtio

   • If the boot virtual file system is initrd, run the following commands:

     lsinitrd /boot/initrd-`uname -r` | grep ` uname -r ` | grep xen

     lsinitrd /boot/initrd-`uname -r` | grep ` uname -r ` | grep virtio

   If the names of the native Xen and KVM drivers are displayed in the command output, the drivers have been installed.

   [root@CTU10000xxxxx home]# lsinitrd /boot/initramfs-`uname -r`.img | grep ` uname -r`| grep xen
   -rwxr--r--   1 root     root        54888 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/block/xen-blkfront.ko 
   -rwxr--r--   1 root     root        45664 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/net/xen-netfront.ko 
    
   [root@CTU10000xxxxx home]# lsinitrd /boot/initramfs-`uname -r`.img | grep ` uname -r`| grep virtio
   -rwxr--r--   1 root     root        23448 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/block/virtio_blk.ko 
   -rwxr--r--   1 root     root        50704 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/net/virtio_net.ko 
   -rwxr--r--   1 root     root        28424 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/scsi/virtio_scsi.ko 
   drwxr-xr-x   2 root     root            0 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/virtio 
   -rwxr--r--   1 root     root        14544 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/virtio/virtio.ko 
   -rwxr--r--   1 root     root        21040 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/virtio/virtio_pci.ko 
   -rwxr--r--   1 root     root        18016 Jul 16 17:53 lib/modules/2.6.32-573.8.1.el6.x86_64/kernel/drivers/virtio/virtio_ring.ko

Make sure that the ECS has been configured successfully, or the ECS may become unavailable after the specifications are modified.

Step 4: Modify Specifications

1. Log in to management console.
2. Click in the upper left corner and select your region and project.
3. Click . Under Compute, click Elastic Cloud Server.
4. On the Elastic Cloud Server page, view the status of the target ECS.

   If the ECS is not in Stopped state, click More in the Operation column and select Stop.

5. Click More in the Operation column and select Modify Specifications.

   The Modify ECS Specifications page is displayed.

6. Select the new ECS type, vCPUs, and memory as prompted.
7. (Optional) Set DeH.

   If the ECS is created on a DeH, you can change the DeH where the ECS resides.

   To do so, select the target DeH from the drop-down list. If no DeH is available in the drop-down list, it indicates that DeH resources are insufficient and cannot be used to create the ECS with specifications modified.

8. Select the check box to confirm that the configuration is complete.
9. Click OK.
   

   • The cloud platform automatically creates a system disk snapshot for you. After the specifications are modified, manually delete the snapshot on the snapshot page if you have verified that services are running properly.
   • If ECS specifications failed to be modified and the ECS becomes unavailable, reinstall the OS. This operation will clear the data on the system disk while the data on data disks is not affected.

(Optional) Step 5: Check Disk Attachment

After a Xen ECS is changed to a KVM ECS, disk attachment may fail. Therefore, check disk attachment after specifications modification. If disks are properly attached, the specifications modification is successful.

• Linux ECS

  For details, see Why Does the Disk Attachment of a Linux ECS Fail After I Modify the ECS Specifications?

Follow-up Procedure

If the ECS specifications have been modified but the OS cannot be started after remote login, contact customer service or reinstall the ECS OS to resolve this issue. For details, see Reinstalling the OS.

Reinstalling the OS will clear the system disk data, but the data on data disks is not affected.

After the specifications are modified, manually delete the snapshot on the snapshot page if you have verified that services are running properly.