|
SDSF uses the HASPINDX data set when building the SYSLOG display. It contains
information related to all SYSLOG jobs and data sets on the spool.
The first time a user accesses the LOG display, SDSF determines which HASPINDX data set to use and allocates it if necessary. The HASPINDX data set can be specified in ISFPARMS with the INDEX and INDXVOL keywords of the ISFPMAC macro or OPTIONS statement.
Alternatively, the data set can be specified by preallocating the ddname HASPINDX prior to accessing the SDSF LOG display. SDSF then uses the data set referenced by that ddname.
Sharing a HASPINDX data set
One or more systems can share a HASPINDX data set. This reduces the number of data sets, but contention and reduced performance may result, depending on SDSF LOG activity. For that reason, IBM does not recommend sharing the HASPINDX data set. Note that HASPINDX data sets are typically small and that sharing of the HASPINDX data set results in no functional, usability or other external difference to end users.
Several configuration options are available for the HASPINDX data set:
- Systems can share a HASPINDX data set. To share the HASPINDX data set, specify the same HASPINDX data set name in ISFPARMS. SDSF will allocate the same data set and it will be shared across the systems. This requires the least number of data sets but may result in degraded performance due to contention.
- Each system can have its own HASPINDX data set. To use unique data sets, specify unique data set names in ISFPARMS. This usually results in improved performance, but several data sets will be needed.
Note that the use of system symbols as part of the data set name may simplify this approach. For example, the OPTIONS statement INDEX(ISF.&SYSNAME..HASPINDX) could result in SDSF allocating HASPINDX data set ISF.SY1.HASPINDX on SY1 and ISF.SY2.HASPINDX on SY2. Support for system symbols was added with OS/390 V2R10 and z/OS V1R1 SDSF.
- Individual users or SDSF batch jobs can have their own HASPINDX data set. To accomplish this, preallocate a data set to ddname HASPINDX. This eliminates contention but many data sets will be needed.
HASPINDX data sets in a mixed environment
During LOG processing, SDSF checks the level of the HASPINDX data set, which is
maintained in a special record within the data set. If the level of the HASPINDX data set is not compatible with the level expected by SDSF, SDSF formats the data set. Formatting involves initializing the data set and preprocessing the syslog data sets on spool.
Formatting time varies based on the size of the HASPINDX data set and the number of syslog jobs and data sets on spool.
With z/OS V1R2 SDSF, the format of the HASPINDX data set changed to accommodate the larger job numbers available in that release. The first reference to a HASPINDX data set at this level (or a new data set) will cause it to be formatted. The data set can then be shared with other z/OS V1R2 SDSF systems because they are all at the same level. However, if the HASPINDX data set is to be shared with other SDSF releases, such as OS/390 R7
through OS/390 R10, or z/OS 1.0, compatibility APARs are required.
The compatibility APARs (which are applicable to FMIDs HQX1900 and HQX7703) cause SDSF to process the HASPINDX data set at the z/OS V1R2 level. These APARs are only relevant if the HASPINDX data set is shared between SDSF at differing release and maintenance levels. These APARs are:
- HQX1900 - PQ44206 (PTF UQ53592)
- HQX7703 - PQ44206 (PTF UQ53593) and PQ48590 (PTF UQ54028)
As a result, you should be aware of several considerations related to sharing of the HASPINDX data set during the transition period when systems are at different release and maintenance levels.
- If the HASPINDX data set is not shared between systems, there is no impact to SDSF LOG regardless of whether the compatibility APARs are installed or not.
- If the HASPINDX data set is shared between z/OS V1R2 SDSF (HQX7705) and either HQX1900 or HQX7703, then the applicable compatibility APAR must be applied to HQX1900 and/or HQX7703.
- The HASPINDX data set can be shared between HQX1900 or HQX7703 systems as long as the compatibility APARs are applied consistently, that is, you must either install the APAR on each system sharing the HASPINDX data set, or install it on none of the systems.
As a result, if you have several pre-z/OS V1R2 systems sharing a single HASPINDX data set and are going to add a z/OS V1R2 system, you need to install the compatibility APARs to each pre-z/OS V1R2 system before it can share the HASPINDX data set with the z/OS V1R2 system. Further, once the compatibility APAR is applied to a pre-z/OS V1R2 system, it can only share a HASPINDX data set with a pre-z/OS V1R2 system on which the compatibility APAR has been applied or with a z/OS V1R2 system.
To accomplish this on a system by system basis, you would simply change the data set name in ISFPARMS on the system on which you are installing the compatibility APAR. This will cause SDSF to allocate a different data set on that system and thus it will not be shared with the other SDSF systems. The APAR can then be installed on that system and everything will work as before. As the compatibility APAR is applied to the remaining systems, you can change ISFPARMS so that the data set is shared between those systems that have the APAR installed.
Once you have installed the APARs to all the systems, you can then go back to a shared HASPINDX data set if desired.
|
