Working in SDSF and handling spool problems

The panels beyond ST

Most people learn SDSF through the status panel. The other panels answer specific questions faster. Which ones you can open depends on your SAF permissions and the SDSF level at your site.

Action characters

You type an action character in the NP column beside a row. The same character means the same thing on most panels.

Many columns can be overtyped: on the I panel you can type a new class or priority over the old one, and SDSF issues the matching JES2 command for you. Every action is checked against your authority. In most shops JESSPOOL profiles let you act on your own jobs but not other people's, and production jobs are restricted to operations and support IDs.

How spool space gets used up

Spool is shared by every job, started task and TSO user on the system, or across the whole sysplex with a multi-access spool. It fills when output is created faster than it is purged:

• A looping program writing DISPLAY lines or messages without end.
• Started tasks that run for months and keep adding to their logs.
• Held output that nobody reads or purges, building up day after day.
• A large dump written to SYSOUT, or many jobs failing and producing big listings at once.

JES2 warns when spool use passes the warning threshold with the $HASP050 message, which names the resource (for example TGS, track groups) and the percentage used. If spool runs out, jobs cannot write output, new work may not be accepted, and the whole system degrades.

$HASP050 JES2 RESOURCE SHORTAGE OF TGS - 85% UTILIZATION REACHED

Relieving a spool shortage

1. Confirm with $D SPOOL or the SDSF SP panel how full the spool is and how fast it is growing.
2. Find the biggest and fastest-growing users. In SDSF you can sort output and job panels by size or record count; a single job with millions of lines is usually the culprit.
3. Stop the growth: if a job is looping, cancel it after agreeing with its owner and support lead.
4. Free space safely: purge old output under the site's retention rules, prioritising test output and output already archived.
5. Do not purge output needed for diagnosis or audit until it has been saved, for example with XDC to a dataset.
6. Afterwards, add controls: output line limits, retention rules, or regular automated purges.

Output limits help prevent repeats. A job or step can set a maximum number of lines (for example LINES= on the JOB statement, or OUTLIM= on a SYSOUT DD). When a limit is exceeded the job can be cancelled with system completion code S722, depending on site settings. That turns a site-wide spool problem into one failed job.