Enterprise load balancing configuration parameters
Use the load balancer configuration parameters to manage enterprise load balancing. To access the load balancing properties in a central manager, go to . In a managed unit, go to .
| Parameter | Default value | Description |
|---|---|---|
| ALLOW_POLICY_MISMATCH_BETWEEN_APPLIANCES | 1 | Whether the load balancer relocates the units such as S-TAP® and universal connectors to a managed unit that has a different policy. Valid values:
|
| APPLIANCE_HW_PROFILE_INDICATORS | NUM_PROCESSORS:CPU_SPEED:CPU_CACHE:CPU_CORES:MEMORY_SIZE |
The load balancer can account for managed units' hardware profile indicators. The load balancer uses the following column names (as a colon-separated list) from the APPLIANCE_RESOURCE_INFO table to evaluate the hardware profile. Valid values: Column names from the APPLIANCE_RESOURCE_INFO table Under normal circumstances, do not change this parameter. |
| DEFAULT_STAP_MAX_CONTRIBUTION_TO_MAX_QUEUE_USAGE | 0.1 |
When units such as S-TAP or universal connector is assigned to a managed unit for the first time, the value of this parameter defines the temporary max load contribution of the unit to the temporary max used queue until the real load is collected from the managed unit (after the interval defined by the TIME_TO_IGNORE_STAP_CONNECTION_RELATED_LOAD parameter). Valid values: 0.1 to 1 in increments of 0.1 Under normal circumstances, do not change this parameter. |
| DEFAULT_STAP_MAX_QUEUE_USAGE | 0.15 |
When units such as S-TAP or universal connector is assigned to a managed unit for the first time, the value of this parameter defines the temporary max used queue until the real load is collected from the managed unit (after the interval defined by the TIME_TO_IGNORE_STAP_CONNECTION_RELATED_LOAD parameter. Valid values: 0.10 to 1 in increments of 0.10 Under normal circumstances, do not change this parameter. |
| DISK_USAGE_ADJUSTMENT_FACTOR | 100 | The percentage of the managed unit's system var disk usage that is compared to the unit
utilization threshold 1. If the disk usage * DISK_USAGE_ADJUSTMENT_FACTOR is
greater than Threshold 1, then the managed unit is considered loaded. For example,
DISK_USAGE_ADJUSTMENT_FACTOR = 50, and the 'utilization thresholds' system var
disk usage threshold 1 = 40. The managed unit is considered loaded if the system var disk usage is
80 or higher, since 50% * 80 meets the threshold of 40. For more details, see Unit utilization and Unit utilization details reports. Valid values: 1 to 200 |
| ENABLE_DYNAMIC_LOAD_COLLECTION | 1 | Controls the load collection method. Valid values:
Changing this parameter triggers an immediate recalculation of the next full load collection time. |
| ENABLE_FAILOVER_GROUPS_REBALANCE | 1 | Controls automatic relocation of units such as S-TAP or universal
connectors from the failover group back to the main managed unit group when a managed unit in the
main managed unit group is available again. Valid values:
|
| ENABLE_RELOCATION | 1 | Relocation of resources (rebalancing) is a process that the load balancer runs after full
load collection. Relocation here means to transfer the units such as S-TAP or universal
connectors from loaded managed units to less loaded manged units. Valid values:
|
| EVALUATE_ADDITIONAL_LOAD_INDICATORS | 0 | Evaluate additional load indicators when you classify a managed unit as loaded. Valid values:
|
| LOAD_BALANCER_ENABLED | 1 |
Controls the load balancer feature. Valid values:
If disabled on the managed unit, the load balancer (running on the central manager) does not collect load information from that managed unit. All the units such as S-TAPs and universal connectors that are connected to that managed unit do not participate in load balancing. If disabled on the central manager, the load balancer service is not available for the units. On the central manager, enabling this parameter (after it was disabled) triggers an immediate full load collection from all the managed units that are enabled for load balancing. |
| LOADED_SNIFFER_QUEUE_USAGE_THRESHOLD | 0.6 |
A managed unit is considered loaded if its sniffer has at least one queue whose size reaches the LOADED_SNIFFER_QUEUE_USAGE_THRESHOLD. Valid values: 0.10 to 1 in increments of 0.10 Under normal circumstances, do not change this parameter. |
| MAX_CONCURRENT_LOAD_COLLECTIONS | 10 |
The maximum number of concurrent load collection processes the load balancer runs at any point in time. That is, the number of concurrent, nonpersistent, remote SQL connections from the central manager to the managed unit. Valid values: 1 to 100 |
| MAX_RELOCATIONS_BETWEEN_FULL_LOAD_COLLECTIONS | 3 |
Defines the maximum number of total relocations (between managed units) allowed after a full load collection. Valid values: ≥-1 Negative values indicate that unlimited relocations are allowed. |
| MAX_RELOCATIONS_PER_MU_BETWEEN_FULL_LOAD_COLLECTIONS | 3 |
The maximum number of unit relocations allowed from a specific managed unit during any one period of full load. For example, for S-TAPs, this parameter is the maximum number of S-TAPs that can be relocated per managed unit. If you have two loading S-TAPs, and the value is set to 1, then only one of these S-TAPs can be moved for a specific managed unit. If the value is set to 0, then S-TAP does not relocate. Valid values: ≥-1 Negative values indicate that unlimited relocations are allowed. |
| REBALANCE_IF_MU_CLASSIFIED_AS_LOADED_N_TIMES_IN_M_HOURS | 1:168 |
Loaded managed units can be rebalanced only if they are classified as loaded for a specified number of instances over a specified period of hours. For example, a value of 1:168 requires that a managed unit is classified as loaded at least 1 time during a period of 168 hours. Valid values: ≥0: ≥0 |
| STATIC_LOAD_COLLECTION_INTERVAL | 720 |
Static managed unit load collection interval (in minutes). If ENABLE_DYNAMIC_LOAD_COLLECTION is set to 0, the load balancer collects the load from all the managed units at the interval that is specified by STATIC_LOAD_COLLECTION_INTERVAL. Valid values: ≥10 |
| TIME_TO_IGNORE_STAP_CONNECTION_RELATED_LOAD | 10 |
When you collect the load statistics for units such as S-TAPs or universal connectors of each managed unit, avoid including data that represents the initial connection of the units to the managed unit. This data can indicate traffic spikes that create a false-positive for the load balancer. This parameter tells the load balancer to ignore the load of the units for the specified number of minutes after the units is connected to the managed unit. Valid values: ≥5 |
| USE_APPLIANCE_HW_PROFILE_FACTOR | 1 |
The load balancer can use managed units' hardware profile indicators (specified by the parameter APPLIANCE_HW_PROFILE_INDICATORS) when it evaluates vacant managed units for relocating units such as S-TAPs or universal connectors. Valid values:
|
| USE_IP_ADDRESS_FORMAT_FOR_ALLOCATIONS | 0 |
Whether the load balancer returns the addresses of allocated managed units such as S-TAPs or universal connectors as a list of hostnames or IP addresses. Valid values:
|