Use this command to set the timeout value for the record cache subsystem
(RCS) state change pending monitor (SCPM) routine. The RCS SCPM
monitors the DASD I/O to ensure that the DASD devices are responding in a
timely method. This time value allows the monitoring of RCS devices
that have entered a state change pending condition because of an error
recovery action at the device or control unit.
This command does the following:
- Starts the RCS SCPM (if the value specified for this command is not
0).
- The RCS SCPM attempts to monitor and report stalled module queues because
of RCS state change pending conditions for DASD devices.
- If the state change pending condition persists for a time period exceeding
the specified interval, the RCS SCPM will initiate action to take the module
offline.
- Note:
- The state change pending condition timing period represents the time period
from the point at which the state change pending unit check is reported by the
device until a valid ending device status is received for a given I/O
operation.
Requirements and Restrictions
You can enter this command only for the basic subsystem (BSS).
Format
- Time-seconds
- specifies the time interval that the RCS SCPM waits before initiating
action to take the device offline once an RCS state change pending condition
is detected, where seconds is the number of seconds from
0-9999.
Notes:
- If you specify 0, no SCPM monitoring will be performed.
- The time interval is initially set to 0 by the TPF system.
Additional Information
- When running on a VM system, the value specified for this command is
doubled.
- Enter the ZSONS DISPLAY SCP command to display the current timeout value
for the RCS SCPM.
- The actual amount of time that the state change pending condition persists
before the SCPM begins to take the device offline is determined by a
combination of the value specified for the TIME parameter and the scan time
frequency interval at which the SCPM runs (the default value is 5
seconds). For example, if the RCS timeout value is specified as 15
seconds and the scan time frequency interval is 5 seconds, the following
timing events occur:
- T0
- The device enters a state change pending condition as a result of unit
check error presentation.
- T1
- Timing of the state change pending condition begins at the next scan time
interval following T0 and is indicated by the CYED0001I online message.
The elapsed time from the recognition of the state change pending condition
(T0) until SCPM timing actually begins is based on the scan time
frequency. Therefore, in this example, the actual elapsed time from T0
to T1 can be 0 to 5 seconds.
- T2
- The next scan determines that the state change pending timing is active
and reports that the condition still exists through the CYED0004I online
message. The actual elapsed time from T0 to T2 can be 5 to 10
seconds.
- T3
- The next scan determines that the state change pending timing is active
and reports that the condition still exists through the CYED0005I online
message. The actual elapsed time from T0 to T3 can be 10 to 15
seconds.
- T4
- The next scan determines that the state change pending timing is still
active. The recovery action to take the device offline is now initiated
because the RCS timeout value (15 seconds in this example) has elapsed since
T1. The actual elapsed time from T0 to T4 can be 15 to 20
seconds.
The previous example showed how the actual state change pending
condition could have been active from 15 to 20 seconds before the recovery
action to take the device offline begins
Examples
In the following example, the state change pending timeout value for the
RCS SCPM recovery routine is set to 15 seconds.
+--------------------------------------------------------------------------------+
|User: ZSONS A SCP T-15 |
| |
|System: SONS0007I 12:22:42 RCS STATE-CHANGE-PENDING TIMEOUT VALUE IS 15 |
+--------------------------------------------------------------------------------+
Related Information