Linux requirements
This topic lists the requirements for a Linux® server to be captured and deployed by IBM® Cloud Infrastructure Center.
For the supported guest Linux distributions and root disk types on z/VM®, refer to Guest operating system support for z/VM
Note:
SLES 15 SP2 cann't be installed directly on FBA disk type on z/VM for some known issue. To make an image using FBA as root disk type for SLES 15 SP2, you need to install SLES 15 SP1 on FBA disk first and upgrade to SLES 15 SP2, then capture this server into an image to be deployed by IBM Cloud Infrastructure Center.
The Linux server that is used as the source of the captured image should meet the following criteria:
When creating images used to deploy virtual machines on ECKD or FBA disk, ECKD or FBA disks should be used as the root disk:
The root file system must not be on a logical volume.
Ensure that the image that is deployed via the IBM Cloud Infrastructure Center matches the disk type that is configured by the
compute_disk_poolof/opt/ibm/icic/config.propertiesin the management node. Either a FBA or a ECKD image can be deployed on one compute node, but not both at the same time. The minidisk on which the root file system resides should be a minidisk of the same type, as required for a subsequent deployment (for example, an ECKD disk image should be captured for a subsequent deployment to an ECKD disk). And the disk should not be a full-pack minidisk, since cylinder 0 on a full-pack minidisks is reserved.The virtual address of the root minidisk on the Linux server should be 0100. Otherwise the deployed virtual machine(VM) fails to start.
The root disk should have a single partition. For SLES 15 SP1 or SP2 server, when configuring the partition size, select the Custom Region option, modify the start block from 768 to 24 to make full use of the root disk.
When creating images used to boot a virtual machine from volume, FC SAN volume should be used as the root disk:
When no LVM (Logical Volume Management) is involved, only one partition that is a standard root partition is supported.
LVM is supported with the following limitations:
Only two partitions are supported where the first one is a standard (non-LVM) partition mounted to / and the second one is a LVM physical volume.
/boot should reside on the / partition, instead of being a separate partition.
When deploying virtual machines from such an image to a larger disk, the LVM volume group is resized automatically. You need to manually extend one or many of the logical volumes after the virtual machine is up. How to extend the logical volume, taking RHEL 7 as an example, refer to Growing Logical Volumes.
The following is an example of the supported disklayout when LVM is involved:
[root@redhat79 ~]# df -h Filesystem Size Used Avail Use% Mounted on devtmpfs 3.9G 0 3.9G 0% /dev tmpfs 3.9G 0 3.9G 0% /dev/shm tmpfs 3.9G 8.4M 3.9G 1% /run tmpfs 3.9G 0 3.9G 0% /sys/fs/cgroup /dev/mapper/mpatha1 3.0G 1.3G 1.7G 44% / /dev/mapper/rootvg-homelv 125M 6.9M 118M 6% /home /dev/mapper/rootvg-tmplv 509M 27M 483M 6% /tmp /dev/mapper/rootvg-optlv 253M 14M 240M 6% /opt /dev/mapper/rootvg-varlv 441M 65M 377M 15% /var /dev/mapper/rootvg-vartmplv 125M 6.9M 118M 6% /var/tmp /dev/mapper/rootvg-varloglv 509M 31M 479M 6% /var/log /dev/mapper/rootvg-varauditlv 61M 3.5M 58M 6% /var/audit tmpfs 788M 0 788M 0% /run/user/0 [root@redhat79 ~]# pvs PV VG Fmt Attr PSize PFree /dev/mapper/mpatha2 rootvg lvm2 a-- <2.00g 0 [root@redhat79 ~]# [root@redhat79 ~]# vgs VG #PV #LV #SN Attr VSize VFree rootvg 1 7 0 wz--n- <2.00g 0 [root@redhat79 ~]# [root@redhat79 ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert homelv rootvg -wi-ao---- 128.00m optlv rootvg -wi-ao---- 256.00m tmplv rootvg -wi-ao---- 512.00m varauditlv rootvg -wi-ao---- 64.00m varloglv rootvg -wi-ao---- 512.00m varlv rootvg -wi-ao---- 444.00m vartmplv rootvg -wi-ao---- 128.00m
Note:
If LVM is also used by the IBM Cloud Infrastructure Center management or compute node itself, ensure a different volume group name is used when creating an image.
IBM Cloud Infrastructure Center supports only deploying an image to a disk larger or equal than the source image’s root disk size. Otherwise it would result in a loss of data. The image’s root disk size can be obtained by:
When the root disk is ECKD or FBA:
hexdump -C -n 64 <image_path>When the root disk is volume:
qemu-img info <image_path>The virtual size in the output represents the size of the root disk. The following is an example for an image made from a server whose root disk size is 3.0G.
[root@tstcomp ~]# qemu-img info rhel77_scsi.qcow2 image: rhel77_scsi.qcow2 file format: qcow2 virtual size: 3.0G (3221225472 bytes) disk size: 1.4G cluster_size: 65536 Format specific information: compat: 1.1 lazy refcounts: false refcount bits: 16 corrupt: false