Setting up ShadowImage Volumes


ShadowImage volume definition

To use ShadowImage volumes you must have the proper features installed on your Hitachi Subsystem, and you must have sufficient DASD capacity within the subsystem to support the newly created ShadowImage volumes. For TrueCopy processing, read/write access to the R-Vol must be enabled. Please contact your Hitachi Vantara Support Team to enable “special option 20 and 190 (Mode20, Mode190)”.

To determine the number of ShadowImage volumes to be created, you must identify all of the online volumes that you intend to use FDRINSTANT and categorize them by type (3380 or 3390) and size (number of cylinders). A ShadowImage volume can be only be paired with an online volume of the same type and equal to or larger size; for use with FDRINSTANT, the size (number of cylinders) should be equal. ShadowImage volumes can be any offline volume within the same Hitachi subsystem as the online volume.

If you have sufficient capacity in your DASD subsystem, you can assign ShadowImage volumes for every online volume involved. If not, you can use a “dynamic pool” of ShadowImage volumes that can be used to create duplicates of online volumes as needed; contact BMC Support for guidance if you want to use a pool.

The device addresses assigned to the ShadowImage volumes must be defined as normal 3380 or 3390 DASD volumes in your I/O configuration (IOCP or HCD). However, as shown later in this chapter, when a ShadowImage volume is detached from its online volume for use with FDRINSTANT it has the same internal volume serial as the online volume. This is normally not a problem since it is offline to the operating system, but it can cause confusion if you IPL your operating system system since the IPL process detects those two volumes that have the same serial and asks the operator which should be left offline (message IEA213A). Obviously, the real online volume should remain online so the operator must respond with the ShadowImage volume address. To avoid the possibility of the operator replying incorrectly, you may want to mark the ShadowImage volumes as OFFLINE in your I/O configuration, so that the operating system does not attempt to access them at IPL time. FDRINSTANT is able access them even though they are offline. However, if you use ShadowImage only with FDRINSTANT for ABR, this is not necessary since ABR will modify the volume label to prevent the volume copy from being varied online.

For example, in HCD, here is the option to mark a device offline to the operating system at IPL:

View Device Parameter / Feature Definition Row 1 of 6
Device number . . : 01E0 Device type . . . : 3390
Generic / VM device type . . . . : 3390


Parameter/ Value Req. Description
Feature
OFFLINE Yes Device considered online or offline at IPL

ShadowImage volumes with FDRINSTANT

Alternately, Hitachi allows you to designate certain devices as “reserved for ShadowImage”. These reserved devices appear to be “not ready” to the operating system unless they are in ShadowImage “split” status. An IPL will not try to access these devices unless they are in split status.

Important

When ABR splits a ShadowImage volume, it modifies the volume label so that IPL will not attempt to bring them online, so when you use FDRINSTANT for ShadowImage with ABR, there is no need to mark them offline in the I/O configuration. But if you use FDRINSTANT with FDR, an IPL may see the duplicate volume serial numbers if you have split the ShadowImage volumes and not yet done the backups.

If the Hitachi subsystem is emulating an IBM 3990 control unit, a ShadowImage volume is paired with an online volume using the TSO command:

CESTPAIR DEVN(X'0C84') PRI(X'0080',30158,X'04') – SEC(X'0080',30158,X'1A')

or the ICKDSF command:

PPRCOPY ESTPAIR DDNAME(DISK1) PRI(X'0080',30158,X'04') – SEC(X'0080',30158,X'1A')

The “triplet” of identifying information in PRI (the online volume) and SEC (the ShadowImage copy) can be obtained for any given Hitachi DASD address by the TSO command CQUERY DEVN(uuuu). Each triplet consists of the subsystem ID (SSID), the control unit serial number, and the channel address of the device.

For ShadowImage, the serial is always the same for PRI and SEC since they are in the same control unit. ShadowImage supports special options (such as Quick Split and Consistent Split) by specifying some special codes in the primary serial number field, consult Hitachi ShadowImage documentation for details. See Quick-Split-Support for details on Quick Split and FDRINSTANT-for-SnapShot-Consistent-Backup-Support for details on Consistent Split.

If the Hitachi subsystem is emulating anything other than an IBM 3990 (such as an IBM 2105 ESS), then additional operands are required to specify the LSS (logical subsystem) containing the primary and secondary volumes. In the TSO command, it is added in the PRI and SEC operands, for example,

CESTPAIR DEVN(X'0C84') PRI(X'0080',30158,X'04',X'01') - SEC(X'0080',30158,X'1A',X'01')

But in the ICKDSF command it is an additional operand:

PPRCOPY ESTPAIR DDNAME(DISK1) PRI(X'0080',30158,X'04') - SEC(X'0080',30158,X'1A') LSS(X'01',X'01')

Example job streams later in this section will show ICKDSF examples including the LSS parameters, but if your subsystem is in 3990 emulation mode, you must omit them.

The ESTPAIR will synchronize the ShadowImage volume with the online volume by copying all data on the online volume to the ShadowImage volume; this may take several minutes. Once the copy is complete, all updates to the online volume are reflected on the ShadowImage volume, the subsystem hardware insures that the ShadowImage volume is always an exact copy of its online volume. While the ShadowImage volume pair is active, in duplex mode, the host cannot directly address the ShadowImage volume.

Since all data on the online volume is mirrored on the ShadowImage volume, including the volume label and the VTOC, VTOCIX, and VVDS, the ShadowImage volume has the same volume serial as the online volume.

At the point in time where a backup or copy of the data on the online volume is desired, the ShadowImage volume can be either split from the online volume by a PPRC SUSPEND command or the pair can be reset to simplex mode by a PPRC DELPAIR command. The ShadowImage volume is an exact copy of the online volume at the time that the split or reset occurred. A split (SUSPEND) of the ShadowImage volume permits it to be re-synchronized with the online volume once the backup is completed. The reset to simplex (DELPAIR) removes any logical hardware association of the ShadowImage volume to the online volume and permits the ShadowImage volume to be used in a different ShadowImage pair once the backup is complete.

For ABR, you must split the ShadowImage volume using the PSPLIT statement of ABR. This allows ABR to identify the address of the ShadowImage volume for later processing.

For FDR, FDRDSF, and FDRCOPY, you split the ShadowImage volume using either the TSO command:

CSUSPEND DEVN(X'0C84') PRI(X'0080',30158,X’04’,X'01') – SEC(X'0080',30158,X’1A’,X’01’)

Or the ICKDSF command:

PPRCOPY SUSPEND DDNAME(DISK1) PRI(X’0080’,30158,X’04’) – SEC(X’0080’,30158,X’1A’) LSS(X’01’,X’01’)

Omit the LSS parameters if the subsystem is in 3990 emulation mode.

Once split or reset is complete, the ShadowImage volume can be accessed by its assigned device address and can be backed up or copied with FDRINSTANT. FDRINSTANT can backup or copy data from the ShadowImage volume even though it is offline and has a serial that duplicates the online volume.

When the backup or copy is complete, you can re-synchronize a split (suspended) ShadowImage volume with the online volume with a PPRC ESTPAIR MODE(RESYNC) command. During the time when the ShadowImage volume pair was split, the DASD Subsystem keeps track of all data updated on the online volume. Re-synchronization copies only the updated data from the online volume to the ShadowImage volume, so it is much faster than a full establish operation. Once the ShadowImage volume is again synchronized with the online volume, it is ready to be used for another FDRINSTANT operation.

For ABR, the step that moves the backup data to tape has an option PPRC=(USE,RET) that will automatically re-synchronize the ShadowImage volume when the backup is complete.

For FDR, FDRDSF, and FDRCOPY, you re-synchronize the ShadowImage volume using either the TSO command:

CESTPAIR DEVN(X'0C84') PRI(X'0080',30158,X'04',X'01') – SEC(X'0080',30158,X'1A',X'01') MODE(RESYNC)

or the ICKDSF command:

PPRCOPY ESTPAIR DDNAME(DISK1) PRI(X'0080',30158,X'04') - SEC(X'0080',30158,X'1A') LLS(X'01',X'01') MODE(RESYNC)

Omit the LSS parameters if the subsystem is in 3990 emulation mode.

Important

The copying of data to ShadowImage volumes is done internally by the Hitachi subsystem and is a low priority task (real I/Os get priority). You may need to wait for the copy to complete before backing up the volume.

You never need to bring a ShadowImage volume online or re-label it or re-initialize it with ICKDSF, as long as it is used only for FDRINSTANT.

Important

If you use the ShadowImage Quick Split function, then you do not need to do an ESTPAIR/CESTPAIR function before the backup. Quick Split allows you to establish and split the ShadowImage volume instantly in a single operation. Likewise, you do not need to do SUSPEND/CSUSPEND (for FDR, FDRDSF, and FDRCOPY. Quick Split allows you to create ShadowImage copies with no previous setup. Details are in Quick-Split-Support.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*