Storage (DASD/disk) considerations
Each zCX instance requires multiple VSAM linear data sets (LDS) to be allocated for its exclusive use. You must define a High Level Qualifier (HLQ) that the provisioning workflow uses to allocate these VSAM LDS data sets. The HLQ value, which is specified in the provisioning workflow, can be multiple qualifiers, but the length must be less than or equal to to 28 characters (including periods). The VSAM LDS are fully allocated during provisioning as primary extents only. Secondary extents are not supported. Consider having unique HLQs for each set of zCX appliance instances that have unique users or purposes, such as PROD, DEV, and TEST.
zCX VSAM data set requirements
- To allow allocation of data sets greater than 4 GB, use a DATACLAS with Extended Format and Extended Addressability.
- Use a MGMTCLAS that specifies the Partial Release attribute as NO to prevent unused space from being released when a data set is closed.
For SMS-managed data sets, attributes such as STORCLAS, DATACLAS, and MGMTCLAS must be assigned either by specifying the values in the provisioning workflow or by SMS Automatic Class Selection (ACS) routines. A valid storage group must be assigned by your ACS routines, and the volumes in that storage group should have the same performance attributes as the STORCLAS assigned to the data set.
For non SMS-managed data sets, a VOLSER must be specified in the provisioning workflow.
zCX VSAM data set permissions
- Each zCX provisioning user needs a security profile which permits the allocation of VSAM data sets.
- Each zCX provisioning user requires ALTER access to the zCX instance’s VSAM data sets.
- Each user ID assigned to a zCX started task will require CONTROL access to the zCX instance’s VSAM data sets.
zFS authority requirements
A Z file system (zFS) is allocated and mounted for each zCX instance. You can use the same HLQ as used for VSAM data set allocation, or specify a different HLQ to allocate the zFS. Primary extents and secondary extents are both supported on zFS allocation. The same DATACLAS, STORCLAS, and MGMTCLAS requirements apply to zFS allocation and VSAM data set allocation.
z/OS provisioning users needs permission to allocate, mount, unmount, and delete a zFS on the zCX
instance directory. Permission is granted by setting up non-privileged user mount
UNIXPRIV class resource, and providing users access to the
SUPERUSER.FILESYS.USERMOUNT resource.
| Data set type | Purpose | Size requirement | Can additional space be added? |
|---|---|---|---|
| Root disk | Linux root file system | Greater than or equal to 4 GB | No, only through deprovisioning and reprovisioning |
| Configuration disk | Holds configuration data for zCX appliance instances | Greater than or equal to 2 MB | No, only through deprovisioning and reprovisioning |
| User Data disk | Holds all docker images, containers, logs, and volumes | Greater than or equal to 20 GB recommended (workload dependent) | Yes, additional disks can be added to the data pool (zCX recycle required) |
| Swap Data disks | Optionally used by Linux kernel for paging/swapping when virtual memory exceeds real memory | If used, greater than or equal to 2 GB (workload dependent) | Yes, additional disks can be added to the data pool (zCX recycle required) |
| Diagnostics and Logs (DLOGS) Data disk | Holds diagnostic data, logs, and FFDC information | Greater than or equal to 1 GB | No, only through deprovisioning and reprovisioning |
| Instance Directory zFS | Holds the zCX appliance image, configuration file, and FFDC information. | Greater than or equal to 4 GB | Can be expanded by secondary extents |
Considerations for allocating large disk sizes for zCX instances
Beginning with OA60920 service level, zCX z/OSMF provision and
add_data_disks workflows provide the capability to allocate large disk sizes up to
1024 GB for zCX instances. Large disks require additional time to format the disks for zCX use.
Formatting of the disks (VSAM linear data sets) is performed during the first start of the zCX
instance. As a result, there will be longer startup times for the first boot of zCX instances when
large disks are specified.