Description: Before z/OS® V1R7,
the HFS file system was the primary hierarchical file system. As of
z/OS V1R7, you can use any combination of HFS and zFS file systems.
Because zFS has higher performance characteristics than HFS and is
the strategic file system, you should migrate your HFS file systems
to zFS.
The HFS and zFS file system types in mount statements and command
operands are now generic file system types that can mean either HFS
or zFS. Based on the data set type, the system will determine which
is appropriate. But note that you must still specify a type (HFS or
zFS and it cannot be defaulted), and if the type you specify is not
correct for the file system being mounted, any associated parameter
string setting in the mount statement or command is ignored, even
though the system sets the type correctly and processes the mount.
Element or feature: |
z/OS UNIX. |
When change was introduced: |
z/OS Distributed File Service became the strategic
file system in z/OS V1R7. |
Applies to migration from: |
z/OS V1R13 and z/OS V1R12. |
Timing: |
Before installing z/OS V2R1. |
Is the migration action required? |
No, but recommended because the action is
planned to become a requirement in a future release. zFS is the strategic
file system for z/OS UNIX and
continues to be enhanced to provide superior performance, reliability,
and data integrity.
z/OS V2R1 will support HFS file systems.
|
Target system hardware requirements: |
None. |
Target system software requirements: |
None. |
Other system (coexistence or fallback) requirements: |
None. |
Restrictions: |
- Understand the zFS recommendations and limits. For more information,
see "Minimum and maximum file sizes" in z/OS Distributed File Service zFS Administration
- The DDNAME() keyword of the BPXPRMxx ROOT and MOUNT statements
is not supported by zFS. Use the FILESYSTEM(name)
keyword instead.
|
System impacts: |
None. |
Related IBM® Health
Checker for z/OS check: |
Use check USS_HFS_DETECTED to verify all file
systems mounted. This check issues exception message BPXH068E if any
HFS file systems are found. |
Steps to take: Follow these steps:
- Before beginning the migration, do the following:
- Establish backout procedures.
- Decide on naming conventions.
- Decide on unavailability.
- Understand any cloning or deployment changes required by zFS systems
being linear data sets. Considerations would include any copy utility
invocations, BPXPRMxx specifications for symbolics, and placement
of zFS file systems on system volumes.
- Perform the conversion from an HFS to zFS file system.
Tip:
Use the BPXWH2Z tool to perform the conversion. It is an ISPF-based
tool that migrates HFS file systems to zFS file systems. Using its
panel interface, you can alter the space allocation, placement, SMS
classes, and data set names. A HELP panel is provided. With this tool,
you can:
- Migrate HFS file systems (both mounted and unmounted) to zFS file
systems. If the HFS being migrated is mounted, the tool automatically
unmounts it and then mounts the new zFS file system on its current
mount point.
- Define zFS aggregates by default to be approximately the same
size as the HFS. The new allocation size can also be increased or
decreased. In z/OS V1R13, ZFS could require up to four times (4X)
the space that HFS did; therefore, for z/OS V2R1 be sure to allocate
more space. See zFS: Accommodate new DASD space requirements.
- Have the migration run in TSO/E foreground or UNIX background.
You can use the JCL sample ISPBTCH in SYS1.SAMPLIB to
invoke BPXWH2Z as an ISPF batch job. Before you run the job, read
the Notes section. When you run BPXWH2Z on your z/OS system,
make sure it uses that same z/OS level of the pax command. You can
manually migrate from an HFS to zFS file system without using the
tool. However, you would need to allocate and format the target zFS
file systems.
You can use the JCL sample ISPBTCH in SYS1.SAMPLIB
to invoke BPXWH2Z as an ISPF batch job. Before you run the job, read
the notes section. When you run BPXWH2Z on your z/OS V1R13 system, it uses the z/OS V1R13 level of the pax command.
This level was enhanced in z/OS V1R7 for sparse file support and other
characteristics that are of concern when migrating from an HFS to
zFS file system. You can manually migrate from an HFS to zFS file
system without using the tool. However, you would need to allocate
and format the target zFS file systems.
Requirement:
The BPXWH2Z tool requires the zFS address space to be operational.
Therefore, before attempting to migrate existing HFS to zFS file systems
using BPXWH2Z, make sure that the zFS address space has been successfully
configured and initialized.
Tip: You can dynamically
migrate the HFS sysplex root in a shared file system configuration
to zFS while the root is in use, without disrupting workloads. Although
the shared file system configuration is required, the sysplex can
be a single system. For the steps involved in migrating the sysplex
root file system from HFS to zFS, see Migrating the sysplex root
file system from HFS to zFS after IPLing z/OS V1R12.
- Change policies and scripts, and so forth, to reflect the change
from the HFS file system to zFS file system. Use the RMF Monitor III
option to report on zFS activity. Refer to migration action "Monitor
the zFS file system activity" for information on gathering zFS activity
in RMF Monitor III.
Tip: Use the RMF Monitor III option
to report on zFS activity.
Note: Beginning in z/OS V2R1, the default
of RMF Monitor III option was changed to NOZFS. So, you need to explicitly
specify ZFS option.
- The DDNAME(name) keyword of the BPXPRMxx ROOT and MOUNT statements
is not supported by zFS. If you use them. change these statements
to use the FILESYSTEM(name) keyword instead.
Migrating the sysplex root file system from HFS
to zFS after IPLing z/OS V1R12:
Before
you begin the migration:
- Ensure that the following requirements have been met:
- All systems in the sysplex are at the V1R12 level.
- The current sysplex root file system PFS, and the new sysplex
root file system PFS, are up in all the systems in shared file system
configuration.
- Be aware of the following restrictions:
- The current sysplex root file system must be mounted as a read-only
file system.
- The systems that do not meet the requirements for this migration
action cannot join the sysplex during the sysplex root file system
migration processing, but they can join the sysplex after the sysplex
root migration has completed.
- The current sysplex root and the new sysplex root must be either
HFS or zFS in any combination. If the new sysplex root is zFS, then
it must be HFS-compatible.
- The sysplex root or any directories on it cannot have been exported
by the DFS or
SMB server.
- Note the following:
- Remote NFS mounts of the sysplex root or any directories on it
are considered active use of the current sysplex root file system.
- During the migration, the new zFS sysplex root file system must
not be HSM-migrated, mounted, or in use.
- Mount parameters are preserved during the migration or replacement
of the sysplex root file system of the same file system type (PFS).
They are dropped if the file system type is different.
- Directories, data, files, and links are not copied from one file
system to another.
Perform the migration as follows:
- Ensure that a file system has been mounted read-only as the current
sysplex root file system. When the root is mounted read-only, there
are no function-shipping clients as long as physical paths to the
DASD are available to each system. To verify that there are no function-shipping
clients, issue:
D OMVS,F,NAME=root_file_system_name
You
should see CLIENT=N on each system.
- Allocate and set up the new zFS sysplex root file system:
- Create a new zFS file system to be used as the new sysplex root
file system. z/OS Distributed File Service zFS Administration discusses
creating and managing zFS file systems.
Rules:
- The UID, GID and the permission bits of the root directory in
the new sysplex root file system must be same as the root directory
in the current sysplex root file system.
- If the SECLABEL class is active and the MLFSOBJ option is active,
the security label for the new zFS file system must match the assumed
security label of the current sysplex root file system.
- On the new sysplex root file system, set up the active mount points
and the symbolic links. The mount points and symbolic links must be
the same as the ones on the current sysplex root file system. You
can set them up either (1) manually or (2) by using the pax shell
command to populate the new sysplex root file system using the existing
sysplex root as a source. To do it manually, create a mount point
in the existing sysplex root (for example, /newroot) and mount the
new sysplex root file system in the MODE(RDWR) on that mount point.
After mounting the new sysplex root file system, manually issue MKDIRs
and ln -s to create the mount point directories and symbolic links
similar to the existing sysplex root file system. Note that the new
sysplex root file system must contain all active mount points and
symbolic links exactly as on the existing sysplex root file system.
- Use the pax shell command to populate
the new file system, using the existing sysplex root as a
source.
- Unmount the new zFS file system.
- On any system in the shared file system configuration, issue:
F OMVS,NEWROOT=new.root.file.system.name,COND=<Yes|No>
- YES
- Proceed conditionally. The system checks for active usage in the
current sysplex root file system and reports the active usage in a
BPXF245I message. If file activity is found, the command fails
with EBUSY return code and JrActivityFound reason code. If file activity
is not found, the command continues processing to replace the sysplex
root. YES is the default.
- NO
- Proceed unconditionally. The system checks for active usage in
the current sysplex root file system and reports the active usage
in a BPXF245I message. Replacement of the sysplex root file
system will continue.
The migration of the sysplex root file system will
begin. During the migration, active connections to files and directories
in the current sysplex root file system are broken. After the migration
completes:
- The root CWD(‘/’) is updated on all systems in the
sysplex to point to the new sysplex root file system.
- New opens go to the new sysplex root file system. The current
sysplex root for the root directory is replaced for all processes
in all systems. The current directory for root directory is replaced
for any processes using it
- Old connections in the previous sysplex root file system might get
EIO errors.
- Update the TYPE parameter and name of the sysplex root
file system in the BPXPRMxx member of SYS1.PARMLIB.
Because the DDNAME() keyword of the BPXPRMxx ROOT and CMOUNT statements
is not supported by zFS, change these statements to use the FILESYSTEM(name)
keyword instead.