Job degradation display commands
Use these commands to display degradation data for address spaces.
- MONJnn
- Displays plot for DEXAN slot number nn or jobname (after
first execution).
- Type:
- Minor of DEX
Executing MONJ a second time results in the following plot:
MONJ01 JOB1 % |0 1 2 3 4 5 6 7 8 9 0| + Using CPU 5.6 |--> . . . . . . . . . .| + Waiting for CPU 33.0 |-------------> . . . . . . .| + Private Page-in Wait 20.4 |--------> . . . . . . . .| + Control Unit 25X Queued 15.0 |------> . . . . . . . . .| + Disk MVS307 735 Active 14.9 |------> . . . . . . . . .|
Note: You may have noticed that the individual wait states shown in this display do not add up to 100%. By default, DEXAN excludes from the display wait reasons that account for less than 5% of the total wait time. You can use the THRS command to adjust this threshold. The MONJ command is also described under Job analysis control commands, because this command can control data collection and display data. - PLTAnn
- Displays a plot of job degradation analysis data for DEXAN slot
number nn.
- Type:
- Minor of DEX
- PLTJnn
- Displays a plot of job degradation
analysis data for DEXAN slot number nn, jobname, or ASID.
- Type:
- Minor of DEX
To display a plot of the most significant wait reasons affecting a job, use the PLTA, PLTJ, or MONJ command.
A sample use of the MONJ command appears earlier in this section. Following is an example of the PLTA and PLTJ command. For example:
PLTJ02 ======= PLTJ T010HP32 ======= PLTA05
In this example, both the DEXAN slot number and jobname were used.
PLTJ02 T063D ASID=279 % |0 1 2 3 4 5 6 7 8 9 0| + Using CPU 5.0 |--> . . . . . . . . . .| + Disk TSO001 253 Queued 30.0 |------------> . . . . . . .| + Private Page-in Wait 20.0 |--------> . . . . . . . .| + Control Unit 25X Queued 15.0 |------> . . . . . . . . .| + Disk MVS307 735 Active 15.0 |------> . . . . . . . . .| + Waiting for CPU 5.0 |--> . . . . . . . . . .| =============================================================================== PLTJ03 T010HP32 ASID=332 % |0 1 2 3 4 5 6 7 8 9 0| + Using CPU 24.5 |---------> . . . . . . . .| + Disk MVWK56 137 Queued 62.1 |------------============>> . . . .| + Waiting for CPU 5.2 |--> . . . . . . . . . .| =============================================================================== PLTA05 SCJ05FL ASID=472 % |0 1 2 3 4 5 6 7 8 9 0| + Using CPU 5.1 |--> . . . . . . . . . .| + Tape Mount Pending 48.8 |------------=======>. . . . . .| + Swapped: Det-Wait 25.4 |----------> . . . . . . . .| + Logical Channel 14 Queued 7.2 |--> . . . . . . . . . .| + Tape 010259 A99 Active 6.4 |--> . . . . . . . . . .|
Using CPU, which is a productive state, always appears first on a plot.
In this example, both job T063D and T010HP32 suffer from severe degradation because of disk I/O contention. Job SCJ05FL waits for almost half of its time on Tape Mounts Pending.
- PCTAnn
- Displays wait reasons by percentage for job nn.
- Type:
- Minor of DEX
Use the PCTA command with the DEXAN slot number to display a percentage listing of wait reasons.
- PCTJnn
- Displays wait reasons by percentage for job nn or jobname.
- Type:
- Minor of DEX
Use PCTJ command to display the wait reasons for a job (address space).
For example:
PCTJ OMEGAMON PCTA5
results in:
PCTJ02 CPU STI + OMEGAMON 0.4 99.6 PCTA05 CPU PAG UNI CPW 4A3 4A0 A9A 4A4 157 Q4A4 140 + PRMS40BW 9.8 28.2 21.7 18.4 9.0 7.9 5.8 4.6 2.1 1.9 .6
This display shows two address spaces: OMEGAMON and PRMS40BW. OMEGAMON shows only STI waits, but PRMS40BW shows degradation due to paging, unilateral swaps, waiting for CPU cycles, and I/O contention.
The first row of the percentage display lists execution states where a workload spends time. The numbers in the second row are the percentage of time that a workload spends in an address space. For a listing of all possible execution states, see Figure 1. For more information on execution states, see Execution state summary.
Only 14 execution states can be displayed at one time. To scroll horizontally to see any additional wait reasons, type 1 in column 1 next to the command to move to the second screen. Type 2 in column 1 to move to the third screen, and so on. To return to the first screen, enter 0 or blank in column 1. Likewise, type 1 in column 1 next to the command to return to the second screen from any screen.
- CNTAnn
- Displays counts of wait reasons for DEXAN slot number nn.
- Type:
- Minor of DEX
- CNTJnn
- Displays counts of wait reasons for DEXAN slot number nn jobname,
or ASID.
- Type:
- Minor of DEX
For example:
CNTJ USERPSP
results in:
CNTJ01 CPU PAG UNI CPW 4A3 4A0 A9A 4A4 QA9A 157 Q4A0 140 + USERPSP --- 307 236 206 98 86 71 50 35 23 21 10
The counts show you the actual results of DEXAN sampling or data collection. Each column heading is an execution state where a job is spending time. The numbers are the actual number of times DEXAN found a job in a particular execution state.
No counts appear for CPU because using CPU is determined by an alternate method. For a complete list of all possible execution states, see Figure 1. For more information about execution states, see Execution state summary. These counts are the raw data from which the resource impact profile, plot, and percentage are drawn. These three displays are what you will rely on most of the time because they are easier to interpret than a count. However, it is good to become familiar with counts because they can help you become familiar with your system and with DEXAN. Also, if you know where the displays come from, they will be even easier to use.
The CNTJ command cannot be used with the IPRO command.