SETCACHE
the out-of-synchronization condition indicates if an internal copy was done or that the out-of-synchronization condition is caused by a probable user error.
2.
Specify only the NOCOPY parameter when both the primary and secondary volumes have been initialized with ICKDSF, using the same initialization parameters, and contain no application data.
3.
Just as the parameter name suggests, nothing is copied from the primary to the secondary.
PACE(n) specifies the number of tracks from 1 to 255 that are to be copied without interruption during the Storage Control copy operation to establish a duplex pair (SETSECONDARY); to establish a duplex pair from a suspended duplex pair (RESETTODUPLEX); or to reestablish a duplex pair from the primary of a suspended pair and a user-specified alternate (REESTABLISHDUPLEX).
Specifying PACE(0) defines an uninterruptible (dedicated) copy operation.
Specifying PACE(1) to PACE (255) defines the number of tracks to copy before releasing the device for any outstanding device activity (when there is no more activity, the copy of the next n tracks resumes). The default is 15 tracks.
Note:
Specify PACE(1) or (2) for optimum device availability. Using PACE(0) or a large PACE value may lock out other activity to the volume for a long time.
Using SETCACHE
Setting Caching On for the Subsystem
SETCACHE SUBSYSTEM ON enables normal caching for the subsystem. If cache storage is disabled or in a pending state, the command fails.
Note:
Pinned data, a cache failure, or cache disabled for maintenance can cause the command to fail.
Abbreviation:
SETC
Setting Caching On for a Device
SETCACHE DEVICE ON sets caching on for individual devices. The prerequisite for device caching is subsystem caching on. This command can fail due to pinned data.
Setting Cache Fast Write On for the Subsystem
SETCACHE CACHEFASTWRITE ON sets cache fast write on for the subsystem.
With cache fast write on for the subsystem, all caching volumes use cache fast write for any channel program specifying it. If cache fast write is not on for the subsystem, the specification is ignored in channel programs that request it.
Prerequisites for cache fast write are: v Subsystem caching on v Device caching on
Setting Nonvolatile Storage On for the Subsystem
SETCACHE NVS ON enables use of the nonvolatile storage for the subsystem and connects the battery to the NVS.
Chapter 31. SETCACHE
363
SETCACHE
If the command fails, it may be due to: v Pinned data v NVS failure v NVS disabled for maintenance.
If the probable cause is pinned data, use LISTDATA PINNED SUBSYSTEM or
DEVSERV
to identify which volumes in the subsystem have pinned data. Fix the problem that is preventing destage, if possible. If the volume cannot be repaired, use DASD installation recovery procedures. If the pinned volumes have DASD fast write active, issue DASDFASTWRITE OFF and then DASDFASTWRITE
PENDINGOFF
to clear the pinned tracks and set DASD fast write off.
DISCARDPINNED can also be used; follow DASD recovery procedures. If the probable cause is NVS failed or disabled, fix the problem and reissue SETCACHE
NVS ON
.
Setting DASD Fast Write On for a Volume
SETCACHE DASDFASTWRITE ON activates DASD fast write for the specified volume. The procedure to activate DASD fast write is:
1.
SETCACHE SUBSYSTEM ON
2.
SETCACHE DEVICE ON
3.
SETCACHE NVS ON
4.
SETCACHE DASDFASTWRITE ON
Creating a Duplex Pair from Two Simplex Volumes
The procedure to create a duplex pair is:
1.
Identify the primary and secondary volumes
2.
Issue SETCACHE SUBSYSTEM ON
3.
Vary the target secondary volume offline to all systems
4.
SETCACHE NVS ON
5.
SETCACHE SETSECONDARY specifying the primary volume id and secondary address
6.
SETCACHE DEVICE ON for the primary volume
7.
(optional) SETCACHE DASDFASTWRITE ON
Note:
1.
If DASD fast write or device caching was on for the primary volume before
SETCACHE SETSECONDARY was issued, DASD fast write or device caching is on for the pair after it is established.
2.
Device caching may be activated to speed up establishing the duplex pair.
3.
When you use dual copy to migrate from 3390 devices to RAMAC devices, an invalid format 4 DSCB is created on the target device. An ICKDSF APAR must be installed to fix the format 4 DSCB. See your IBM representative for the latest service level.
4.
If either volume was previously acquired by an LPAR and has not been released, an error message is issued with text that states that the duplex pair could not be established because path-groups are not compatible.
Changing 3990 and 9390 Cache and NVS Operating Modes
This discussion covers the various Storage Control SETCACHE commands and the resulting actions. The commands presented here are not arranged in any priority and operate independently of one another.
364
z/OS DFSMS Access Method Services Commands
SETCACHE
Setting Cache Off for the Subsystem
SETCACHE SUBSYSTEM OFF sets normal caching off for the subsystem. The following actions occur: v Device caching, cache fast write, and DASD fast write stop. Device status, with respect to each of these functions, is retained and the active functions resume when subsystem caching is set on. When DASD fast write and dual copy are set on together (fast dual copy), the data is destaged from NVS to both devices.
v Each duplex pair is set to suspended state on the first write operation to the pair. Also, any duplex pairs with out-of-synchronization cylinders are suspended after SETCACHE SUBSYSTEM OFF is issued. However, if there is no write activity to a duplex pair, and all cylinders are in synchronization, the pair is not set to suspended duplex state.
v The Storage Control destages all modified data from cache and NVS to DASD.
Setting Caching Off for a Device
SETCACHE DEVICE OFF sets device caching off for the specified volume. Cache fast write and DASD fast write operations stop for that device. Modified data for the volume is destaged to the DASD. The cache fast write and DASD fast write volume status is retained so that when cache is set on again, these functions resume.
Setting Nonvolatile Storage Off for the Subsystem
SETCACHE NVS OFF deactivates the NVS. This command also disconnects the
NVS battery on a 3990 Model 6, or 9390, or on a 3990 Model 3 with either RPQ
8B0174 or 8B0175 installed. DASD fast write is stopped and all modified DASD fast write data is destaged to the appropriate volumes. Cache fast write data is not destaged. Dual copy changed cylinder logging is terminated for dual copy volumes, but both copies of dual copy volumes continue to be updated.
Either a hardware failure in a duplex pair while the NVS is deactivated, or a utility power outage, or a loss of cache, or the cache being set off at this time causes all duplex pairs to be set to suspended duplex state at the next write I/O to each pair.
Setting DASD Fast Write Off for a Volume
SETCACHE DASDFASTWRITE OFF sets DASD fast write off for the designated volume. All DASD fast write data for the specified volume is destaged.
Setting Cache Fast Write Off for the Subsystem
SETCACHE CACHEFASTWRITE OFF sets cache fast write off for the logical
DASD subsystem. All cache fast write data is destaged to the DASD.
Resetting a Duplex Pair to Two Simplex Volumes
SETCACHE RESETTOSIMPLEX terminates a duplex pair and restores both volumes to simplex operations. The primary volume retains the DASD fast write status of the dual copy logical volume. The secondary volume assumes the device status defaults, unless the pair was suspended, in which case the caching status is inactive. Normally, all updates in NVS and cache are destaged to both the primary and the secondary volumes.
The primary volume contains copies of all updates to the volume. Because of a
DASD fast write pending condition (probably due to another job), modified tracks might not have been destaged to the secondary volume. If the volumes must be
Chapter 31. SETCACHE
365
SETCACHE
identical, either use a utility program to compare the two volumes, or change the secondary volume serial number using ICKDSF, vary it online, and copy the primary volume to the secondary volume.
Destaging All Modified Data in Cache and NVS to DASD
SETCACHE DESTAGE specifies that all modified data in cache and NVS is scheduled for destage to DASD. This command is usually issued for shutdown and is included in the operator command Halt End of Day (Halt EOD).
Resetting All Cache and NVS Operating States to Defaults
SETCACHE REINITIALIZE performs the following: v Resets all Storage Control cache and NVS operating states to the initial installation subsystem defaults (in addition all status tracks are reinitialized).
v
Resets all status information to default state.
v Discards all modified tracks in the cache or NVS. All dual copy logical volumes are set to simplex state and all dual copy logical volumes are reset to simplex.
To save the modified data, issue SETCACHE DESTAGE before issuing
SETCACHE REINITIALIZE.
Because of the nature of this command, all volumes in the subsystem must be varied offline to all attached system images before the command is issued.
Attention:
Carefully consider the use of SETCACHE REINITIALIZE before issuing the command and ensure that all data has been destaged to DASD. This command resets all dual copy logical pairs to simplex state and resets the
CCA/DDC pointers to their original state.
Using Dual Copy to Migrate Volumes
The following steps show how to migrate data from one device to another without disrupting the application. Separate procedures are provided for simplex volumes and duplex pair volumes.
Note:
1.
When you use dual copy to migrate volumes, remember that the correspondence between the original CCA and DDC addresses is switched and that the Storage Control retains the updated CCA-to-DDC address correspondence.
Use the message response to the DEVSERV PATHS command to identify the current CCA-to-DDC relationship.
2.
When you use dual copy to migrate from 3390 devices to RAMAC devices, an invalid format 4 DSCB is created on the target device. Use ICKDSF to fix the format 4 DSCB. See your IBM representative for the latest service level of
ICKDSF that provides this function.
Migrating Simplex Volume
The following procedure is used to migrate simplex volumes, that is, volumes that are not already part of a duplex pair. This procedure is valid for any two devices qualified to be part of a duplex pair.
1.
Vary the target secondary volume offline to all attached systems.
2.
Issue SETCACHE SETSECONDARY.
3.
After the copy completes, issue SETCACHE SUSPENDPRIMARY.
366
z/OS DFSMS Access Method Services Commands
SETCACHE
4.
Issue SETCACHE RESETTOSIMPLEX.
5.
The original device is no longer in use. The application continues to access the same UCB (MVS device address), but is using a different physical device.
6.
Proceed with required activities. Use ICKDSF to change the volser if required.
Note:
When you use dual copy to migrate from 3390 devices to RAMAC devices, an invalid format 4 DSCB is created on the target device. Use ICKDSF to fix the format 4 DSCB. See your IBM representative for the latest service level of ICKDSF that provides this function.
If DASD fast write was active before this procedure was executed, it remains active after the procedure completes.
1.
Before the duplex pair is established, the channel connection address (CCA), which is known to MVS, and the director to device connection (DDC) address, which is the physical address known within the subsystem, are the same as when initialized.
2.
After creating the duplex pair, both addresses directly correlate.
3.
When SETCACHE SUSPENDPRIMARY is issued, the Storage Control swaps the CCA-to-DDC assignments between the primary and the secondary volumes.
4.
The system application continues to use the UCB for device 201. However, the
Storage Control directs the I/O to the physical device addressed by DDC 12
(the target volume). The UCB for the offline device 212 still points to CCA 12, but the Storage Control points CCA 12 to the suspended device (DDC 01).
5.
When SETCACHE RESETTOSIMPLEX is issued, the Storage Control maintains the swapped CCA-to-DDC pointers.
Migrating Duplex Volumes
This procedure is valid for migrating a primary or secondary volume in a duplex pair to another like device.
v Vary the target volume offline to all systems.
v Issue SETCACHE SUSPENDSECONDARY or SUSPENDPRIMARY, depending on whether the secondary or the primary, respectively, needs to be taken out of the duplex pair.
v Issue SETCACHE REESTABLISHDUPLEX with the third volume.
v After the copy completes, perform the required action for the volume suspended from the duplex pair.
Note:
When you use dual copy to migrate from 3390 devices to RAMAC devices, an invalid format 4 DSCB is created on the target device. Use ICKDSF to fix the format 4 DSCB. See your IBM representative for the latest service level of ICKDSF that provides this function.
1.
When the SETCACHE SUSPENDPRIMARY completes, the system device number points to the CCA address of the old primary volume. Internally, the
Storage Control has swapped the CCA-to-DDC pointers. The old primary CCA now points to the old secondary physical device DDC. Thus, the old secondary volume has become the new primary volume; all system I/O is directed to the new primary. The CCA for the old secondary volume now points to the DDC address of the old primary device.
2.
After the REESTABLISHDUPLEX completes, the new secondary is incorporated into the duplex pair, the old primary volume is no longer in the pair, and the
Chapter 31. SETCACHE
367
SETCACHE
CCA of the old secondary now points to the DDC address of the old primary.
To run ICKDSF or any other program against the old primary volume, the old primary should be addressed as 212.
Dual Copy Device Address Mapping
In a dual copy environment, normal subsystem activities can modify the internal mapping of the system address (device number) to the actual device. In IDCAMS reports and in the output of the DEVSERV command, the low-order two digits of the system address are shown as the CCA, whereas the physical device address is shown as the DDC. The subsystem manages the integrity of this mapping, so it need not be a cause for concern. Circumstances exist, however, when the user needs to be aware that dual copy address mapping may have changed: v When device maintenance is required v
When devices are being removed or repositioned v When subsystem re-initialization is required v When it is necessary to vary a device offline for reconfiguration or maintenance, use the CCA to determine the system address.
You can see the current map of system and device addresses by issuing the
DEVSERV command or by running IDCAMS LISTDATA DSTATUS to any device or volume in the subsystem. Along with EREP (for 3380s) and SIM messages (for
3390s), this information may be useful to the service representative for identifying a failing device.
Note:
Because of the asynchronous nature of IDCAMS operations, it is possible that during certain operations, the results from a DEVSERV command could differ from the results of an IDCAMS report. If the two operations start together, the
DEVSERV results are more current.
368
z/OS DFSMS Access Method Services Commands
Chapter 32. SHCDS
Use the SHCDS command to list SMSVSAM recovery associated with subsystems spheres and to control that recovery. This command works both in batch and in the
TSO/E foreground. The functions include the following subcommands: v List subcommands v Subcommands that enable you to take action on work that was shunted v Subcommands to control a manual forward recovery in the absence of a forward recovery utility that supports SMSVSAM protocols v
Subcommands that enable you to run critical non-RLS batch window work when it is not possible to first close out all outstanding SMSVSAM recovery v A subcommand that allows for a subsystem cold start
Recommendation:
After a cold start, if recovery was not completed for any data sets, they are most likely left in a damaged state and must be recovered manually. If the data sets are forward recoverable, their forward recovery logs might also be damaged. Manually recover the data sets (without using forward recovery), take backups of them and of any other data sets that use the forward recovery log, and then delete and redefine the forward recovery log.
Use this command cautiously. See z/OS DFSMSdfp Storage Administration for details
about administering VSAM RLS. See Appendix C, “Interpreting SHCDS Output
Listings,” on page 451 for SHCDS output listings.
The syntax of the access method services SHCDS command is:
Command
SHCDS
Parameters
[LISTDS(base-cluster){JOBS}]|
[LISTSHUNTED{(base-cluster|ALL)]|
[LISTSTAT(base-cluster)]
[LISTSUBSYS(subsystem|ALL)]|
[LISTSUBSYSDS(subsystem|ALL)]|
[LISTRECOVERY(base-cluster)|
[LISTALL]|
[FRSETRR(
base-cluster)]|
[FRUNBIND(base-cluster)]|
[FRBIND(base-cluster)]|
[FRRESETRR(base-cluster)]|
[FRDELETEUNBOUNDLOCKS(base-cluster)]|
[PERMITNONRLSUPDATE(base-cluster)]|
[DENYNONRLSUPDATE(
base-cluster)]|
[REMOVESUBSYS(subsystem)]|
[CFREPAIR(
{INFILE(ddname)|INDATASET(dsname)}
[({LIST|NOLIST})]]|
CFREPAIRDS({
base_cluster_name|partially_qualified_cluster_name})
[CFRESET({INFILE(ddname)|INDATASET(dsname)}]
[({LIST|NOLIST})]]|
CFRESETDS({
base_cluster_name|partially_qualified_cluster_name})
CFQUIRSDS({
base_cluster_name |partially_qualified_cluster_name})
[PURGE{SPHERE(base-cluster)|URID(urid)}]|
[RETRY{SPHERE(base-cluster)|URID(urid)}]|
[OUTFILE(ddname)]
© Copyright IBM Corp. 1973, 2017
369
