- ABNORMALTERM=option
- Specifies
if the abnormal terminations, that are reported by the WLM report
service IWMRPT for a server, are included when computing the weight
of the routing service IWM4SRSC that is called for the server. The
syntax is:
ABNORMALTERM=NO means that the abnormal terminations,
that are reported by the IWMRPT service, are not included when computing
the weight of the IWM4SRSC routing service. ABNORMALTERM=YES means
that the abnormal terminations are included when computing the weight
of the IWM4SRSC routing service.
Values: YES or NO
Default
Value: YES
- BLWLINTHD=option
- Specifies
the threshold time interval for which a blocked address space or enclave
must wait before being considered for promotion.
If the CPU utilization
of a system is at 100%, workloads with low importance (low dispatch
priority) might not get dispatched anymore. This can lead to problems
if the low priority work holds a resource that is required by high
priority workloads. Therefore, if an address space or enclave has
ready-to-run work units (TCBs or SRBs) but does not get CPU service
for the specified time interval because of its low dispatch priority,
it will be temporarily promoted to a higher dispatch priority. Address
spaces that are swapped out are not considered for promotion.
Value range: 1-65535 seconds
Default
Value: 20 seconds
- BLWLTRPCT=option
- Specifies
how much of the CPU capacity is to be used to promote blocked workloads.
This parameter does not influence the amount of CPU service that a
single blocked address space or enclave is given. Instead, this parameter
influences how many different address spaces or enclaves can be promoted
at the same point in time. If the value specified with this parameter
is not large enough, blocked workloads might need to wait longer than
the time interval defined by BLWLINTHD.
Value Range: 0-200
(up to 20%. 0% indicates blocked workload support
is not enabled.)
Default Value: 5
(0.5%)
- CCCAWMT=xxxxxxx
- Specifies
whether to activate or deactivate Alternate Wait Management (AWM).
If AWM is active, SRM and LPAR cooperate to reduce low utilization
effects and overhead. In an LPAR, some n-way environments with little
work appear to require more capacity than expected because of the
time spent waking up idle logical and physical processors to compete
for individual pieces of work. If AWM is active, SRM and LPAR will
reduce this unproductive use of processor so that capacity planning
is more accurate and CPU overhead is reduced.
Value
Range: For HIPERDISPATCH=NO, any value between 1 and 499999 makes
AWM active, and any value between 500000 and 1000000 makes AWM inactive.
AWM is active or inactive only for general CPs, zAAPs, and zIIPs.
For
HIPERDISPATCH=YES, the valid range is 1600 to 3200. Any other value
will be reset to the default value of 3200. AWM is always active and
cannot be turned off.
Default Value: AWM
is active. For HIPERDISPATCH=NO, the default is 12000 (12 ms). For
HIPERDISPATCH=YES, the default is 3200 (3.2 ms).
- CCCSIGUR=option
- Specifies
the minimum mean-time-to-wait threshold value for heavy CPU users.
This constant is used to determine the range of mean-time-to-wait
values which are assigned to each of the ten mean-time-to-wait dispatching
priorities. The specified real time value is adjusted by relative
processor speed to become SRM time to insure consistent SRM control
across various processors.
Value Range: 0-32767 milliseconds
Default
Value: 45
- CNTCLIST=option
- Specifies
if the individual commands in a TSO/E CLIST are treated as separate
commands for transaction control.
In the syntax,
option is either YES or NO. CNTCLIST=NO specifies that the CLIST is
treated as a single transaction. CNTCLIST=YES specifies that each
command is to be treated as an individual transaction. By specifying
CNTCLIST=YES, SRM control of a TSO/E command becomes the same whether
the command is executed explicitly or as part of a CLIST.
Values:
YES or NO
Default Value: NO
- CPENABLE=(xxx,yyy)
- Specifies
the low (ICCTPILO) and high (ICCTPIHI) threshold values for the percentage
of I/O interruptions to be processed through the test pending interrupt
(TPI) instruction path in IOS. SRM uses the following thresholds to
control the number of processors enabled for I/O interruptions.
Value Range |
Basic Mode Default |
LPAR Mode Default |
a=0-100% |
10 |
0 |
b=0-100% |
30 |
0 |
Note: For MVS™ running
in LPAR mode with dedicated CPUs, specify the basic mode default values
for CPENABLE. For recommendations on setting CPENABLE parameters,
please see ATS Flash1033 at the following URL:
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/FLASH10337
- DVIO=option
- Specifies whether
directed VIO is to be active in the system or not.
DVIO=YES, the
default, specifies that directed VIO is to be active in the system;
that is, the NONVIO keyword of the IEASYSxx parmlib member is honored.
DVIO=NO specifies that directed VIO is not to be active in the system;
the NONVIO parameter of the IEASYSxx parmlib member is ignored.
Values:
YES or NO
Default Value: YES
- FULLPRESYSTEM= [YES | NO]
- Specifies the SRM to allow full preemption
for the system address spaces.
In FULLPRESYSTEM=YES mode, SRM allows
full preemption handling for system address spaces. To free up a processor
immediately when such an address space becomes ready, the system issues
signal processor (SIGP) instructions. The SIGP instructions also run
in HiperDispatch mode, so when HIPERDISPATCH=YES and FULLPRESYSTEM=YES,
too much LPAR overhead can result. Therefore:
- Use FULLPRESYSTEM=NO when HIPERDISPATCH=YES
- Use FULLPRESYSTEM=YES when HIPERDISPATCH=NO.
FULLPRESYSTEM=NO Specifies
the SRM to turn off full preemption mode for the system address spaces.
Value
Range: YES or NO
Default Value:
- YES when HiperDispatch is not in effect.
- NO when HiperDispatch is in effect.
- ERV=xxxxxx
-
- Specifies
the number of CPU service units that an address space or enclave is
allowed to absorb when it is possibly causing enqueue contention.
During this “enqueue residency” time,
the address space (including the address space associated with an
enclave) is not considered for swap-out based on recommendation value
analysis. The address space or enclave runs with a high enough priority
to guarantee the needed CPU time.
ERV is in effect for an address
space or enclave that meets one of the following criteria:
- The address space or enclave is enqueued on a system resource
needed by another address space.
- An authorized program in the address space or enclave obtains
control of the resource (even if another address space does not need
that resource) as a result of issuing a reserve for a DASD device
which is SHARED.
Note: SRM determines the execution time equivalent to the specified
ERV by multiplying the ERV by the model-dependent time needed to accumulate
1 CPU service unit.
Example: ERV=2
In
the example above, if an address space consumes 1 service unit in
10 milliseconds, it will be allowed to execute for 20 milliseconds
before it will be eligible for swap-out while enqueued on a resource
requested by other address spaces.
Value Range: 0-999999
Default Value: 500
- HIPERDISPATCH=YES|NO
-
- YES
- Specifies that SRM should turn
on HiperDispatch mode.
- NO
- Specifies that SRM should turn off HiperDispatch
mode.
Note: All partitions with greater than 64 logical
processors defined at IPL are forced to run with HIPERDISPATCH=YES.
After IPL, LPARs with greater than 64 logical processors defined are
unable to switch into HIPERDISPATCH=NO. Also, when HIPERDISPATCH=YES
(when it is supported by the processor), VARY CPU management is automatically
turned off no matter what is specified for VARYCPU.
Note: On
IBM z13™ and
follow-on processors that support multi threading mode, when partitions
specify LOADxx PROCVIEW=CORE, they are forced to run with HIPERDISPATCH=YES.
After IPL, these partitions are unable to switch into HIPERDISPATCH=NO.
Also, when HIPERDISPATCH=YES (when it is supported by the processor),
VARY CPU management is automatically turned off no matter what is
specified for VARYCPU.
Default Value:
For IBM zEnterprise® 196 (and follow-on processors):
YES
For IBM System z10 processors: NO
- IFAHONORPRIORITY=YES|NO
-
- YES
- Specifies that standard processors run both zAAP
processor eligible and non-zAAP processor eligible work in priority
order when the zAAP processors indicate the need for help from standard
processors. The need for help is determined by the alternate wait
management (AWM) function of SRM for both standard and zAAP processors.
Standard processors help each other and standard processors can also
help zAAP processors if YES is in effect. Specifying YES does not
mean the priorities will always be honored because the system manages
dispatching priorities based on the goals provided in the WLM service
definition. AWM should not be disabled when IFAHONORPRIORITY=YES is
in effect. See the description for parameter for a description of AWM.
If
zAAP processors are defined to the LPAR but are not online, the zAAP
processor eligible work units are processed by standard processors
in priority order. The system ignores the IFAHONORPRIORITY parameter
in this case and handles the work as if it had no eligibility to zAAP
processors. The zAAP processor eligible processor times are reported
in RMF™ and SMF for planning purposes.
IBM suggests that you specify or
default to IFAHONORPRIORITY=YES.
- NO
- Specifies that standard processors will never examine zAAP processor eligible work. Note that standard processors also run zAAP processor
eligible work, if it's necessary to resolve contention for resources
with non-zAAP processor eligible work.
Default Value: YES
- IIPHONORPRIORITY=YES|NO
-
- YES
- Specifies that standard processors run both zIIP processor eligible
and non-zIIP processor eligible work in priority order when the zIIP
processors indicate the need for help from standard processors. The
need for help is determined by the alternate wait management (AWM)
function of SRM for both standard and zIIP processors. Standard processors
help each other and standard processors can also help zIIP processors
if YES is in effect, which is the default. Specifying YES does not
mean the priorities will always be honored because the system manages
dispatching priorities based on the goals provided in the WLM service
definition. AWM should not be disabled when IIPHONORPRIORITY=YES is
in effect. See the description for parameter for a description of AWM.
If
zIIP processors are defined to the LPAR but are not online, the zIIP
processor eligible work units are processed by standard processors
in priority order. The system ignores the IIPHONORPRIORITY parameter
in this case and handles the work as if it had no eligibility to zIIP
processors. The zIIP processor eligible processor times are reported
in RMF and SMF for planning purposes.
IBM suggests that you specify or
default to IIPHONORPRIORITY=YES.
- NO
- Specifies that standard processors will never examine
zIIP processor eligible work. Note that standard
processors also run zIIP processor eligible work, if it's necessary
to resolve contention for resources with non-zIIP processor eligible
work.
Default Value: YES
- INITIMP=option
- Specifies
the dispatching priority for JES, APPC, and OMVS initiators. The option
is specified as one of the following values:
An INTIMP value affects only the system on which
it has been set.
Example:
INITIMP=2
With INITIMP=2, if there are service classes with
an importance of 2 and the CPU critical attribute, the dispatching
priority of initiator address spaces will always be lower than the
dispatching priority of the work running in those service classes.
Because CPU critical work with a higher importance level will always
have a higher dispatching priority, the dispatching priority of the
initiator will also be lower than any CPU critical work with an importance
of 1.
With INITIMP=2 and CPU critical work with an importance
of 1 and 3, the initiator will always have a lower dispatching priority
than the work running in the service class with the CPU critical attribute
and an importance of 1 but not for the work running in the CPU critical
service class with an importance of 3.
With INITIMP=2 and no
service classes with the CPU critical attribute and an importance
of 1 or 2, the initiator dispatching priority is set to the enqueue
promotion dispatching priority which is dynamically calculated. Please
note that the initiators will not run at a dispatching priority of
254 as that is only the case if you do not specify the INITIMP parameter
or if you have set it to 0.
Default Value:
0
- IRA405I=n
- Specifies
the percentage of the fixed storage that causes the system to issue
message IRA405I. Specify a value for n to
indicate the storage area, where n can be:
- 0
- Real storage area below 16M
- 1
- Real storage area between 16M and 2G
- 2
- Total real storage
Note that when you specify the percentage value,
you must specify the value without the percent symbol (%), as shown
in the following example:
IRA405I(1)=60
When
each value of the
n is specified, the value
ranges and the default values for the IRA405I parameter can be:
- IRA405I(0)
- Value Range: 0 to 100.
Default Value: 70.
- IRA405I(1)
- Value Range: 0 to 100.
Default Value: 50.
- IRA405I(2)
- Value Range: 0 to 100.
Default Value: 50.
- MANAGENONENCLAVEWORK=option
- Specifies
whether non-enclave transaction work of queue servers and enclave
servers is to be managed or not, where option is
YES or NO.
MANAGENONENCLAVEWORK=YES indicates that SRM is to manage
the non-enclave transaction work towards the first service class period
of the address space performance goal. Based on this expanded performance
management, be sure to verify the performance goals for the service
class of the address spaces that process the enclave work. For a detailed
description, see the chapter "Performance Management of Address Spaces
with Enclaves" in z/OS MVS Programming: Workload Management Services.
MANAGENONENCLAVEWORK=NO
indicates that you expect no work consuming significant CPU service
to be running in the address space outside of an enclave. The CPU
consumption of work running outside of enclaves is not included when
workload management assesses the impact of CPU adjustments for the
enclave work.
This performance management is available for z/OS® V1R12 and later releases.
Note: To
complete a MANAGENONENCLAVEWORK status change from YES to NO or NO
to YES for an active system, SRM issues a policy refresh for that
active system. Workload Manager indicates the policy refresh for this
status change with message IWM065I.
Values: YES or
NO
Default Value: NO
Example:
MANAGENONENCLAVEWORK=YES
In the example, SRM is to manage the
non-enclave work of enclave servers and queue servers
- MAXPROMOTETIME=xxxx
- Specifies
the time during which a resource holder is allowed to run promoted
when it is possibly causing a resource contention. A resource holder
is either an address space or enclave. During this "resource contention
residency" time, the resource holder runs with the highest priority
of all resource waiters to guarantee the needed importance. Also during
this interval, the address space (including the address space associated
with an enclave) is not considered for swap-out based on recommendation
value analysis. If MAXPROMOTETIME=0 is specified,
SRM does not promote any resource holder.
MAXPROMOTETIME is in
effect if the resource manager has notified the address space or enclave
under contention by using the WLM IWMCNTN notification service. See z/OS MVS Programming: Workload Management Services for a description of the WLM IWMCNTN service.
The value specified for MAXPROMOTETIME is multiplied
by 10. This is the time span in seconds during which the address space
or enclave under contention is allowed to run promoted.
Example:
MAXPROMOTETIME=12
With MAXPROMOTETIME=12, SRM
allows an address space or enclave to run 12 × 10
seconds (=2 minutes) promoted. When the
time is exceeded, the promotion is canceled.
Value
Range: 0-1000
Default Value: 6
- MCCAFCTH=(lowvalue,okvalue)
- Specifies the low and the OK threshold values for storage. The lowvalue indicates
the number of frames on the available frame queue when stealing begins.
The okvalue indicates the number of frames
on the available frame queue when stealing ends. SRM
will automatically adjust the actual threshold values based on measurements
of storage usage, but threshold values should never fall below the
values specified in MCCAFCTH. You do not have to specify an MCCAFCTH
value.
If the lowvalue or okvalue is
below the default value, the default value is enforced.
- Value
- lowvalue=1-32767 frames
- Default
- The maximum of 400 and 0.2% of the pageable storage
in the system
- Value
- okvalue=1-32767 frames
- Default
- The maximum of 400 and 0.2% of the pageable storage
in the system
- MCCFXEPR=xxx
- Specifies
the percentage of storage that is fixed within the first 16 megabytes.
SRM uses this threshold to determine when a shortage of pageable storage
exists because there are too many fixed pages.
Value Range:
0-100 percent
Default Value: 92 percent
- MCCFXTPR=xxx
- Specifies
the percentage of online storage that might be fixed.
SRM uses this threshold to determine when a shortage of pageable storage
exists.
Note: SRM uses the lesser of the values, (MCCFXTPR ×
amount of online storage) and (MCCFXEPR × amount of storage
that is fixed within the first 16 megabytes) to set the threshold
frame count so that it can detect a shortage of pageable storage that
is caused by too much page fixing. In this way, SRM can detect a shortage
of pageable storage that is caused by too much page fixing before
or at the same time as a shortage caused by too much paging.
Value range: 0-100 percent
Note: If 100% - MCCFXTPR * the total amount of online frames
is greater than 64 gigabytes, the MCCFXTPR keyword is not used in
determining the threshold at which a shortage of pageable storage
exists.
Instead, on systems with more than 320 gigabytes of storage,
a pageable storage shortage is detected when less than 64 gigabytes
of online storage is pageable.
When calculating the number of
frames that can be page fixed before a pageable storage shortage is
detected, SRM uses the maximum of MCCFXTPR * the total online frames
and total online storage minus 64 gigabytes.
- MT_CP_MODE=1
- MT_ZIIP_MODE=1|2
- Specifies the multithreading (MT)
mode, which is the number of active threads per core for all online
cores with a core type of standard CP, or System z Integrated Information
Processor (zIIP).
Default Value: 1
- MT_CP_MODE=
- The number of active threads for each online CP core. For CPs only a value of 1 is valid.
- MT_ZIIP_MODE
- The number of active threads for each online zIIP core. The value
is limited to what the underlying hardware and operating system support.
A value of 1 is always accepted.
Activating
an MT mode greater than 1, requires the following:
- The hardware must support multithreading for the respective processor
class, namely zIIPs.
- On the HMC Customize/Delete Activation Profiles task, "Do not
end the time slice if a partition enters a wait state" must not be
checked. This is the recommended default setting.
- The system must be IPLed with LOADxx PROCVIEW CORE in effect.
- HIPERDISPATCH=YES must be in effect.
If these requirements are met, the MT mode for all cores
of the respective processor class is changed to the specified value/mode if the current MT mode is different
from that value. In case of a successful change of the MT Mode, message
IWM066 is issued. For an unsuccessful request, message IWM067 is issued.
Changing the MT Mode for all cores of the same type
can take multiple seconds to complete.
- PROJECTCPU=[YES | NO]
- Specifies whether to activate or deactivate the projection of
how work could be offloaded from regular CPs to special assist processors
like the System z Application
Assist Processor (zAAP) and the System
z Integrated Information Processor (zIIP).
Note: - The PROJECTCPU parameter is not necessary if
you can define a zIIP or zAAP as a reserved processor in the LPAR
configuration. The PROJECTCPU parameter is provided for users of earlier
processors on which defining a zIIP as a reserved processor is not
supported when no zIIP is purchased for planning purposes.
- If the installation requires ZIIP statistics but
there is no ZIIP defined (reserved or actual) on the LPAR, then PROJECTCPU=YES
is required regardless of hardware model.
- Any work that is eligible for being offloaded to a special assist
processor will be reported as Special_Processor_on_CP work. This information
can be used to understand the benefit of adding a special processor
into the configuration.
Values: YES or NO
Default Value: NO
- projection will not be done, unless a special assist processor is
configured to the system. In that case, the system will collect projection
values.
- RCCFXET=option
- SRM uses
these thresholds to determine whether the system MPL needs to be increased
or decreased based on the first 16 MB. The values that you can specify
for option are as follows:
- (aaa,bbb)
- Specifies the low (RCCFXETL) and the high (RCCFXETH) percentages
of storage that is fixed within the first 16 megabytes.
Value range: 0-100 percent
- RCCFXTT=option
- SRM uses
these thresholds to determine if the system MPL needs to be increased
or decreased. The values that you can specify for option are
as follows:
- (aaa,bbb)
- Specifies the low (RCCFXTTL) and the high (RCCFXTTH) percentages
of online storage that is fixed.
Value range: 0-100 percent
- RMPTTOM=xxxxxx
- Specifies
the SRM invocation interval. The specified real-time interval is adjusted
by relative processor speed to become SRM time in order to ensure
consistent SRM control across various processors. The relationship
of real time to SRM time for each processor is described in the “Advanced
SRM Parameter Concepts” topic of z/OS MVS Initialization and Tuning Guide.
Value Range: 1000-999999 msec
Default
Value:
- 3000 (for systems with a uni-processor speed of more than 100
MIPS)
- 1000 (for systems with a uni-processor speed of 100 MIPS or less)
- RTPIFACTOR=xxx
- Specifies
how much the server performance index (PI) should affect the server
routing weights returned by WLM services IWM4SRSC and IWMSRSRS with
FUNCTION=SPECIFIC.
- When RTPIFACTOR is 0, the server weight is independent from the
server PI.
- When RTPIFACTOR is 100 and the server PI is bigger than 1, the
server weight is divided by the server PI.
- When RTPIFACTOR is between 1 and 99, it results in a corresponding
intermediate influence of the server PI on the server weight returned
by WLM.
Value Range: 0-100
Default Value: 100
- STORAGENSWDP=option
- Specifies if the system should select non-swappable
address spaces to resolve a storage shortage.
STORAGENSWDP=YES
specifies that the system should also select non-swappable address
spaces to resolve the
storage shortage. The
system then sets non-swappable address spaces non-dispachable and
issues message IRA210E or IRA410E.
Note: The system
does not set address spaces in service class SYSTEM non-dispatchable.
Values:
YES or NO
Default Value: YES
- STORAGESERVERMGT=option
- Specifies
whether SRM should pass service class importance and goal information
to the storage I/O priority manager in the IBM System Storage® DS8000® series.
The passed information enables the storage I/O priority manager to
provide favored processing for I/O requests of important z/OS workloads that are missing their goals.
STORAGESERVERMGT=YES specifies that SRM should pass
service class importance and goal information to the storage I/O priority
manager. Before specifying STORAGESERVERMGT=YES, verify that your IBM System Storage DS8000 model incorporates the storage I/O
priority manager feature.
If STORAGESERVERMGT=YES
is specified, the storage I/O priority manager may throttle I/O requests
to facilitate a favored access to storage server resources for other
I/O requests. The storage I/O priority manager considers the properties
of the service class period associated with an I/O request to determine
whether the I/O request should receive a favored processing or may
be throttled:
- For service class periods with a response time goal, the goal
achievement and specified importance are considered.
- For service class periods with a velocity goal, the specified
velocity goal and importance are considered.
- I/O requests associated with the system-provided service classes
SYSTEM, SYSSTC, or SYSSTC1 - SYSSTC5 are not managed by the I/O priority
manager.
- I/O requests associated with service class periods that have a
discretionary goal will always be throttled.
Throttle delays introduced by the storage
IO priority manager are reflected in control unit queue delays. Therefore,
if STORAGESERVERMGT=YES is specified, control unit queue delays are
not considered when the achieved velocity and performance index are
calculated for service class periods with a velocity goal.
If you have significant control unit queue delays in
your installation, you may have to adjust the velocity goal of service
class periods when you specify STORAGESERVERMGT=YES.
Values:
YES or NO
Default Value: NO
- STORAGEWTOR=option
- Specifies
how the system handles address spaces in a critical storage shortage.
STORAGEWTOR=YES specifies that the system presents a
list of address spaces on the console. The operator can reply to a
WTOR (IRA221D or IRA421D) request and select which address space to
cancel. STORAGEWTOR=AUTO is similar to STORAGEWTOR=YES, except that
the IRA210I or IRA420I message presents up to 20 address spaces at
once. This option is useful, when an automation product needs to answer
the WTOR request.
Values: YES, AUTO,
or NO
Default Value: YES
- TIMESLICES=option
- Specifies
the number of timeslices that a CPU-intensive address space or enclave
with a discretionary goal should be given before a dispatchable unit
of equal importance is dispatched.
Increasing this parameter might
increase the processor delay for some CPU-intensive work, but decrease
the number of context switches between equal priority work, and therefore
increase the throughput of the system. This parameter only affects
discretionary work that is CPU-intensive as determined by significant
mean time to wait (MTTW) (see parameter CCCSIGUR).
Value
Range: 1-255
Default Value: 1
- VARYCPU=option
- Specifies
whether LPAR Vary CPU management is available or not available. Note that Vary CPU management is available
only in a partition that is enabled for LPAR weight management. Also,
Vary CPU management is automatically turned off no matter what is
specified for VARYCPU, if HIPERDISPATCH is on.
VARYCPU=YES,
the default, specifies that LPAR Vary CPU Management is available
for this system. VARYCPU=NO specifies that LPAR Vary CPU Management
is not available.
Value Range: YES or NO
Default
Value: YES
- VARYCPUMIN=nn
- Specifies
the minimum number of CPs (nn) that must
stay online during WLM LPAR management. WLM LPAR management will not
take CPs offline below this threshold.
Value Range: 1 - 64
Default
Value: 1
- WASROUTINGLEVEL=0|1
- Specifies
the routing algorithm used by the WLM Websphere routing services.
- 0
- WLM uses the most advanced routing algorithm that
is supported by all systems in the sysplex.
- If the release level of the lowest system is z/OS R9 (or above), routing decisions are based
on available or dispatchable capacity of standard and assist processors.
- If the release level of the lowest system is below z/OS R9, but has the APAR OA16486 installed,
routing decisions are based on available or dispatchable capacity
of standard processors only.
- If the release level of the lowest system is below z/OS R9 and runs without the APAR OA16486 installed,
the routing algorithm round robin is used.
- 1
- WLM uses the routing algorithm round robin. If this option is used, set it on all systems of the
sysplex so that all are using the same algorithm; otherwise, the WebSphere® routing service
on each system uses the specific algorithm setting for the system,
which can lead to inconsistent results.
Default Value: 0
- ZAAPAWMT=xxxxxx
- Specifies an Alternate Wait Management (AWM) value for IBM System z Application Assist Processors (zAAPs)
to minimize SRM and LPAR low utilization effects and overhead. In
an LPAR, some n-way environments with a small workload may appear
to have little capacity remaining because of the time spent waking
up idle zAAPs to compete for individual pieces of work. The ZAAPAWMT
parameter allows you to reduce this time so that capacity planning
is more accurate and CPU overhead is reduced, even though it might
take longer until arriving work gets dispatched.
Value
Range: For HIPERDISPATCH=NO, the valid range is 1 to 499999 microseconds.
Any value between 1 and 499999 makes AWM active. To inactivate AWM,
set CCCAWMT to any value between 500000 and 1000000.
For
HIPERDISPATCH=YES, the valid range is 1600 to 499999 microseconds.
Any other value will be reset to the default value of 3200. AWM is
always active and cannot be turned off.
Default
Value: AWM is active. For HIPERDISPATCH=NO, the default is 12000
(12 ms), and for HIPERDISPATCH=YES, the default is 3200 (3.2 ms).
- ZIIPAWMT=xxxxxx
- Specifies an Alternate Wait Management (AWM) value for the IBM System z9 Integrated Information Processor
(zIIP) to minimize SRM and LPAR low utilization effects and overhead.
In an LPAR, some n-way environments with a small workload may appear
to have little capacity remaining because of the time spent waking
up idle zIIPs to compete for individual pieces of work. The ZIIPAWMT
parameter allows you to reduce this time so that capacity planning
is more accurate and CPU overhead is reduced, even though it might
take longer until arriving work gets dispatched.
Value
Range: For HIPERDISPATCH=NO, the valid range is between 1 and
499999 microseconds. Any value between 1 to 499999 makes AWM active.
To inactivate AWM, set CCCAWMT to any value between 500000 and 1000000.
For HIPERDISPATCH=YES, the valid range is 1600 to 499999
microseconds. Any other value will be reset to the default value
of 3200. AWM is always active and cannot be turned off.
Default Value: AWM is active. For HIPERDISPATCH=NO,
the default is 12000 (12 ms), and for HIPERDISPATCH=YES, the default
is 3200 (3.2 ms).
Note: The
ZAAPAWMT and ZIIPAWMT parameters
internally affect the frequency with which the specialty engines will
check the need for help. If help is required, the zAAP or zIIP processor
signals a waiting zAAP or zIIP to help. When all zAAP or zIIP processors
are busy and IFAHONORPRIORITY/IIPHONORPRIORITY=YES, the zAAP or zIIP
processors ask for help from the standard processors. All available
speciality engines (that is, all zAAPs or all zIIPs) must be busy
before help is asked of the standard processors. Even with IFAHONORPRORITY=YES
or IIPHONORPRIORITY=YES parameters set, the general CP will not help
the specialty engine for work in a discretionary service class.
Reducing
the value specified for ZAAPAWMT or ZIIPAWMT causes
the specialty engines to request help after being busy for a shorter
period of time. If IFAHONORPRIORITY or IIPHONORPRIORITY is set to
YES, help is provided to one CP at a time, in the priority order of
zAAP or zIIP processor eligible work, non-zAAP or non-zIIP processor
eligible work. Reducing the ZAAPAWMT or ZIIPAWMT value
too low can cause the standard processors to run an excessive amount
of zAAP or zIIP processor eligible workload, which might result in
lower priority non-zAAP/zIIP processor eligible work to be delayed.
Conversely, increasing the value specified for ZAAPAWMT
or ZIIPAWMT causes the specialty engines to request help only
after being busy for a longer period of time, which might delay the
standard processors from providing help when it is necessary.