Retrieving Collection Broker Configuration Information

The collection-broker-get function (known as CollectionBrokerGet in the SOAP API) returns an XML collection-broker-configuration node that contains configuration information for the Collection Broker. In C#, a call to this function looks like the following:


    CollectionBrokerGet cbget = new CollectionBrokerGet();
    CollectionBrokerGetResponse cbgetresp = port.CollectionBrokerGet(cbget);
    collectionbrokerconfiguration config = cbgetresp.collectionbrokerconfiguration;

The collection-broker-configuration node can contain the following elements:

always-allow-one-collection - (Boolean, default value true) When set to true, at least one collection is allowed. This prevents pathological cases where no progress can be made. This option takes precedence over the maximum-collections setting.
check-memory-usage-time - (Double, default value 3) The number of seconds to wait between successive checks to determine if a collection needs to be stopped in order to maintain a healthy amount of available memory on the system.
check-online-time - (Double, default value 10) The number of seconds to wait between successive checks of each online collection to determine if it is still online.
crawler-minimum - (Unsigned long, default value 367001600) The minimum number of bytes of memory that a crawler is expected to use. If the calculated memory usage of the crawler is less than this number, the calculated memory usage is set to this value.
crawler-overhead - (Unsigned long, default value 262144000) This number is subtracted from the crawler memory usage as reported by the system. This can be used to adjust memory calculations because the system might report a larger virtual memory usage than the process uses.
current-status-time - (Double, default value 20) The number of seconds to wait between successive checks of each collection to determine the current collection state.
find-new-collections-time - (Double, default value 30) The number of seconds to wait between successive checks to determine if new collections have been added to the system.
indexer-minimum - (Unsigned long, default value 367001600) The minimum number of bytes of memory that an indexer is expected to use. If the calculated memory usage of the indexer is less than this number, the calculated memory usage is set to this value.
indexer-overhead - (Unsigned long, default value 262144000) This number is subtracted from the indexer memory usage as reported by the system. This can be used to adjust memory calculations because the system might report a larger virtual memory usage than the process uses.
live-ping-probability - (Double, default values 0.1) The probability that a collection will be actively pinged. Doing an active ping causes the collection service to be started, which ensures that the collection status is completely up to date. A high probability is not required, as the status is usually be up to date.
maximum-collections - (Integer, default value -1) The maximum number of collections to have running at one time, regardless of allocatable memory. A value of -1 means no limit is applied. Valid values for this configuration value are -1, 0, and any positive integer. To completely disable collections from being started, you must also set the always-allow-one-collection setting to false.
memory-granularity - (Unsigned long, default value 10485760) In order for two collections to be considered to have different memory usages, there must be a difference of this many bytes of memory between the collections. A smaller value creates more distinction between services with similar memory usage.
minimum-free-memory - (Unsigned long, default value 262,144,000 - 250 MB) The number of bytes of memory to always leave free on the system. This prevents the collection broker from completely exhausting available memory.
overcommit-factor - (Double, default value 0.75) A multiplication factor that is applied to the amount of free memory on the system, and used to manage system estimates of the amount of free memory on different systems. Specifying a value of less than one causes the Collection Broker to leave some amount of memory free by allocating only that percentage of available memory. Specifying a value that is greater than 1 enables the Collection Broker to allocate more memory than is physically available, causing the system on which it is running to swap or page.
persistent-save-time - (Double, default value 10) The number of seconds to wait between subsequent saves of persistent status data to the disk.
prefer-requests - (Enumerated string type, possible values search and enqueue, default value search) Select which type of request should be preferred when multiple requests are received, but not enough resources are available to handle all requests.
Note: IBM strongly recommends not changing the value of this option from its default value of search. Changing the value of this option to enqueue when URLs are being enqueued or are present in an offline queue will prioritize those collections over collections that are being used to return search results, which can have a negative impact on the responsiveness of a search application from the user's point of view.
start-offline-time - (Double, default value 15) The number of seconds to wait between successive checks of each collection with data in the offline queue to determine if that collection can be started to process its offline queue.
start-stop-delta - (Unsigned long, 209,715,200 - 200 MB) An internal constant that represents the average memory requirements that are associated with starting or stopping a search collection. This parameter is internal, and cannot be changed.
time-granularity - (Long, default value 30) In order for two collections to be considered to have different idle times, there must be a difference of this many seconds between the collections. A smaller value creates more distinction between services with similar idle times.