The problem indicator table contains examples of indicators. Some problems might need to be investigated using more than one diagnostic procedure to find the cause. If there are several indicators, look for the earliest problem that caused the other problems.
For example, you find several abends and a wait state, look for the earliest abend code and begin diagnosis there.
Problem type | Indicator | System action | System programmer action |
---|---|---|---|
Abend See Diagnosing an abend. |
SVC dump taken and a record of the error (in Logrec) | Produces dump | Review dump to determine if further diagnosis is required |
Message received indicating a system or user abend | Produces record of error | Review response of system message to determine the impact of the abend on the installation. | |
An ABEND dump is produced | Continue processing | Review the dump to determine if further diagnosis is required. | |
|
System actions are the same as the
indicators listed previously. Note: The system might also initiate
recovery actions. See SYSLOG and component trace to determine what
these recovery action were. Some recovery actions can cause data to
be lost, requiring the installation to resubmit jobs or transactions.
|
|
|
Job hang/wait or loop | Job does not end, no further output is produced, and the job can or cannot be CANCEL'ed or FORCE'd | No response | Use the DUMP command to obtain an SVC dump of the hung job. If the DUMP is not successful, consider taking a stand-alone dump. |
System hang or wait | Disabled wait indicated on the HMC and wait state message issued | The system issues a wait state message and loads a disable wait state PSW. The system might load the following into the PSW: X'070E0000 00000000' | Take a stand-alone dump. |
Many jobs are hung in the system | Resource contention | Enter the DISPLAY GRS,C command to check for ENQ
resource and latch contention and take a dump of the holder of the
resource including SDATA=GRSQ. Note: Use the DISPLAY GRS,ANALYZE command
to aid in the discovery of blockers in the system.
|
|
No response to system or subsystem commands entered | No response | Partition the system from the sysplex and take a stand-alone dump. | |
Loop | High processor resource being consumed locking out other work; Excessive spin detected with IEE178I or ABEND071 issued, or both. | ABEND071 issued in an attempt to stop the looping program | Use an online monitor, such as Resource Measurement Facility™ RMF™ or IBM® OMEGAMON® z/OS® Management Console, to determine whether the problem originates from a high priority job in normal processing or from a problem. |
A job is using a high percentage of central processor storage | Processing degrades | Use an online monitor, such as RMF, to determine whether the problem originates from a high priority job in normal processing or from a problem. | |
Enabled wait or performance degradation | System processing slows. | Processing degrades | Use an online monitor, such as RMF, to determine where the problem originates. |
There is a series of WAIT messages followed by a burst of activity | Processing continues | Use an online monitor, such as RMF, to determine where the bottleneck is occurring. | |
Output problem | Job output is missing or is incorrect. | Processing continues | Use GTF or SLIP to trace input and output. |