Module org.snmp4j
Package org.snmp4j.util

Class TableUtils

java.lang.Object
org.snmp4j.util.AbstractSnmpUtility
org.snmp4j.util.TableUtils
public class TableUtils extends AbstractSnmpUtility
The TableUtils class provides utility functions to retrieve SNMP tabular data.
Since:
1.0.2
Version:
2.5.11
Author:
Frank Fock

  • Field Details

  • Constructor Details

    • TableUtils

      public TableUtils(Session snmpSession, PDUFactory pduFactory)
      Creates a TableUtils instance. The created instance is thread safe as long as the supplied Session and PDUFactory are thread safe.
      Parameters:
      snmpSession - a SNMP Session instance.
      pduFactory - a PDUFactory instance that creates the PDU that are used by this instance to retrieve table data using GETBULK/GETNEXT operations.

  • Method Details

    • getTable

      public List<TableEvent> getTable(Target<?> target, OID[] columnOIDs, OID lowerBoundIndex, OID upperBoundIndex)
      Gets synchronously SNMP tabular data from one or more tables. The data is returned row-by-row as a list of TableEvent instances. Each instance represents a row (or an error condition). Besides the target agent, the OIDs of the columnar objects have to be specified for which instances should be retrieved. With a lower bound index and an upper bound index, the result set can be narrowed to improve performance. This method can be executed concurrently by multiple threads.
      Parameters:
      target - a Target instance.
      columnOIDs - an array of OIDs of the columnar objects whose instances should be retrieved. The columnar objects may belong to different tables. Typically, they belong to tables that share a common index or sub-index prefix. Note: The result of this method is not defined if instance OIDs are supplied in this array!
      lowerBoundIndex - an optional parameter that specifies the lower bound index. If not null, all returned rows have an index greater than lowerBoundIndex.
      upperBoundIndex - an optional parameter that specifies the upper bound index. If not null, all returned rows have an index less or equal than upperBoundIndex.
      Returns:
      a List of TableEvent instances. Each instance represents successfully retrieved row or an error condition. Error conditions (any status other than RetrievalEvent.STATUS_OK) may only appear at the last element of the list.

    • getTable

      public List<TableEvent> getTable(Target<?> target, OID[] columnOIDs, OID lowerBoundIndex, OID upperBoundIndex, long timeoutSeconds)
      Gets synchronously SNMP tabular data from one or more tables. The data is returned row-by-row as a list of TableEvent instances. Each instance represents a row (or an error condition). Besides the target agent, the OIDs of the columnar objects have to be specified for which instances should be retrieved. With a lower bound index and an upper bound index, the result set can be narrowed to improve performance. This method can be executed concurrently by multiple threads.
      Parameters:
      target - a Target instance.
      columnOIDs - an array of OIDs of the columnar objects whose instances should be retrieved. The columnar objects may belong to different tables. Typically, they belong to tables that share a common index or sub-index prefix. Note: The result of this method is not defined if instance OIDs are supplied in this array!
      lowerBoundIndex - an optional parameter that specifies the lower bound index. If not null, all returned rows have an index greater than lowerBoundIndex.
      upperBoundIndex - an optional parameter that specifies the upper bound index. If not null, all returned rows have an index less or equal than upperBoundIndex.
      timeoutSeconds - the maximum number of seconds to wait for the whole table to be retrieved. Set it to 0, to wait forever.
      Returns:
      a List of TableEvent instances. Each instance represents successfully retrieved row or an error condition. Error conditions (any status other than RetrievalEvent.STATUS_OK) may only appear at the last element of the list.
      Since:
      3.4.5

    • createTableRequest

      protected TableUtils.TableRequest createTableRequest(Target<?> target, OID[] columnOIDs, TableListener listener, Object userObject, OID lowerBoundIndex, OID upperBoundIndex, TableUtils.SparseTableMode sparseTableMode)

    • getTable

      public void getTable(Target<?> target, OID[] columnOIDs, TableListener listener, Object userObject, OID lowerBoundIndex, OID upperBoundIndex, TableUtils.SparseTableMode sparseTableMode)
      Gets SNMP tabular data from one or more tables. The data is returned asynchronously row-by-row through a supplied callback. Besides the target agent, the OIDs of the columnar objects have to be specified for which instances should be retrieved. With a lower bound index and an upper bound index, the result set can be narrowed to improve performance.

      This method may call the TableListener.finished(org.snmp4j.util.TableEvent) method before it returns. If you want to synchronize the main thread with the finishing of the table retrieval, follow this pattern: 

            synchronized (this) {
         TableListener myListener = ... {
            private boolean finished;

            public boolean isFinished() {
              return finished;
            }

            public void finished(TableEvent event) {
               ..
               finished = true;
               synchronized (event.getUserObject()) {
                  event.getUserObject().notifyAll();
               }
            }
         };
         tableUtil.getTable(..,..,myListener,this,..,..);
         while (!myListener.isFinished()) {
           wait();
         }
      }
      Parameters:
      target - a Target instance.
      columnOIDs - an array of OIDs of the columnar objects whose instances should be retrieved. The columnar objects may belong to different tables. Typically, they belong to tables that share a common index or sub-index prefix. Note: The result of this method is not defined if instance OIDs are supplied in this array!
      listener - a TableListener that is called with TableEvent objects when an error occured, new rows have been retrieved, or when the table has been retrieved completely.
      userObject - an user object that is transparently supplied to the above call back.
      lowerBoundIndex - an optional parameter that specifies the lower bound index. If not null, all returned rows have an index greater than lowerBoundIndex.
      upperBoundIndex - an optional parameter that specifies the upper bound index. If not null, all returned rows have an index less or equal than upperBoundIndex.
      sparseTableMode - defines how rows with non-existing column values should be handled. Such rows can occur when new rows are being created or rows are removed from an agent while it is being

    • getTable

      public void getTable(Target<?> target, OID[] columnOIDs, TableListener listener, Object userObject, OID lowerBoundIndex, OID upperBoundIndex)
      Gets SNMP tabular data from one or more tables. The data is returned asynchronously row-by-row through a supplied callback. Besides the target agent, the OIDs of the columnar objects have to be specified for which instances should be retrieved. With a lower bound index and an upper bound index, the result set can be narrowed to improve performance.

      This method may call the TableListener.finished(org.snmp4j.util.TableEvent) method before it returns. If you want to synchronize the main thread with the finishing of the table retrieval, follow this pattern: 

            synchronized (this) {
         TableListener myListener = ... {
            private boolean finished;

            public boolean isFinished() {
              return finished;
            }

            public void finished(TableEvent event) {
               ..
               finished = true;
               synchronized (event.getUserObject()) {
                  event.getUserObject().notifyAll();
               }
            }
         };
         tableUtil.getTable(..,..,myListener,this,..,..);
         while (!myListener.isFinished()) {
           wait();
         }
      }
      Parameters:
      target - a Target instance.
      columnOIDs - an array of OIDs of the columnar objects whose instances should be retrieved. The columnar objects may belong to different tables. Typically, they belong to tables that share a common index or sub-index prefix. Note: The result of this method is not defined if instance OIDs are supplied in this array!
      listener - a TableListener that is called with TableEvent objects when an error occurred, new rows have been retrieved, or when the table has been retrieved completely.
      userObject - an user object that is transparently supplied to the above call back.
      lowerBoundIndex - an optional parameter that specifies the lower bound index. If not null, all returned rows have an index greater than lowerBoundIndex.
      upperBoundIndex - an optional parameter that specifies the upper bound index. If not null, all returned rows have an index less or equal than upperBoundIndex.
      Since:
      1.5

    • getDenseTable

      public void getDenseTable(Target<?> target, OID[] columnOIDs, TableListener listener, Object userObject, OID lowerBoundIndex, OID upperBoundIndex)
      Gets SNMP tabular data from one or more tables. The data is returned asynchronously row-by-row through a supplied callback. Besides the target agent, the OIDs of the columnar objects have to be specified for which instances should be retrieved. With a lower bound index and an upper bound index, the result set can be narrowed to improve performance.

      This implementation must not be used with sparse tables, because it is optimized for dense tables and will not return correct results for sparse tables.

      Rows that appear or disappear while being retrieved, are dropped and will be not returned partially (see TableUtils.SparseTableMode.denseTableDropIncompleteRows).
      Parameters:
      target - a Target instance.
      columnOIDs - an array of OIDs of the columnar objects whose instances should be retrieved. The columnar objects may belong to different tables. Typically they belong to tables that share a common index or sub-index prefix. Note: The result of this method is not defined if instance OIDs are supplied in this array!
      listener - a TableListener that is called with TableEvent objects when an error occurred, new rows have been retrieved, or when the table has been retrieved completely.
      userObject - an user object that is transparently supplied to the above call back.
      lowerBoundIndex - an optional parameter that specifies the lower bound index. If not null, all returned rows have an index greater than lowerBoundIndex.
      upperBoundIndex - an optional parameter that specifies the upper bound index. If not null, all returned rows have an index less or equal than lowerBoundIndex.
      Since:
      3.0

    • getMaxNumRowsPerPDU

      public int getMaxNumRowsPerPDU()
      Gets the maximum number of rows that will be retrieved per SNMP GETBULK request.
      Returns:
      an integer greater than zero that specifies the maximum number of rows to retrieve per SNMP GETBULK operation.

    • setMaxNumRowsPerPDU

      public void setMaxNumRowsPerPDU(int numberOfRowsPerChunk)
      Sets the maximum number of rows that will be retrieved per SNMP GETBULK request. The default is 10.
      Parameters:
      numberOfRowsPerChunk - an integer greater than zero that specifies the maximum number of rows to retrieve per SNMP GETBULK operation.

    • getMaxNumColumnsPerPDU

      public int getMaxNumColumnsPerPDU()
      Gets the maximum number of columns that will be retrieved per SNMP GETNEXT or GETBULK request.
      Returns:
      an integer greater than zero that specifies the maximum columns of rows to retrieve per SNMP GETNEXT or GETBULK operation.

    • setMaxNumColumnsPerPDU

      public void setMaxNumColumnsPerPDU(int numberOfColumnsPerChunk)
      Sets the maximum number of columns that will be retrieved per SNMP GETNEXT or GETBULK request. The default is 10.
      Parameters:
      numberOfColumnsPerChunk - an integer greater than zero that specifies the maximum columns of rows to retrieve per SNMP GETNEXT or GETBULK operation.

    • isSendColumnPDUsMultiThreaded

      public boolean isSendColumnPDUsMultiThreaded()

    • setSendColumnPDUsMultiThreaded

      public void setSendColumnPDUsMultiThreaded(boolean sendColumnPDUsMultiThreaded)
      Enable multi-threaded column PDU sending. If set to true and if the maxNumColumnsPerPDU value is less than the number of columns to be retrieved in a TableUtils request, then the requests for the columns will be splitted in two or more columns and the requests will be send to the agent concurrently without waiting for the response of the first/previous PDU. By default, this is disabled.
      Parameters:
      sendColumnPDUsMultiThreaded - if true, multi-threaded processing of column PDUs is enabled, otherwise only a single request will be sent to the agent on behalf a getTable(Target, OID[], OID, OID) or getTable(Target, OID[], TableListener, Object, OID, OID).

    • isCheckLexicographicOrdering

      public boolean isCheckLexicographicOrdering()
      Indicates whether a single request on behalf of getTable(Target, OID[], OID, OID) or getTable(Target, OID[], TableListener, Object, OID, OID) is sent to the agent or not.
      Returns:
      false if single requests are sent, true if more than a single request may be sent at a time.

    • getIgnoreMaxLexicographicRowOrderingErrors

      public int getIgnoreMaxLexicographicRowOrderingErrors()
      Gets the maximum number of rows with wrong lexicographic ordering whicb will be return on a table retrieval with isCheckLexicographicOrdering() set to true.
      Returns:
      the number of ignored row ordering errors.
      Since:
      2.5.11

    • setIgnoreMaxLexicographicRowOrderingErrors

      public void setIgnoreMaxLexicographicRowOrderingErrors(int ignoreMaxLexicographicRowOrderingErrors)
      Sets the maximum number of rows that will be returned with status RetrievalEvent.STATUS_WRONG_ORDER before the table retrieval will be stopped. If this value is set to zero and lexicographic ordering check is enabled by setCheckLexicographicOrdering(boolean), then table retrieval finishes immediately when the error is detected. Otherwise, retrieval continues until the maximum errors are detected and then the row cache will be returned too, although it may contain already incomplete rows based on correctly or incorrectly (!) ordered rows. The default value is three. That means, even if the ordering error occurs at the end of the table and
      Parameters:
      ignoreMaxLexicographicRowOrderingErrors - the maximum numbers of rows with lexicographic ordering error to be returned before finishing table retrieve automatically. Setting this value has no effect if isCheckLexicographicOrdering() is false.
      Since:
      2.5.11

    • setCheckLexicographicOrdering

      public void setCheckLexicographicOrdering(boolean checkLexicographicOrdering)
      Enables or disables lexicographic ordering checks. By default, those checks are enabled, because otherwise with agents, that do not implement correct lexicographic ordering, endless looping could occur.
      Parameters:
      checkLexicographicOrdering - false to disable checks which could increase performance.
      Since:
      2.5.10

    • createRow

      public <A extends Address> ResponseEvent<A> createRow(Target<A> target, OID rowStatusColumnOID, OID rowIndex, VariableBinding[] values)
      Creates an SNMP table row for a table that supports the RowStatus mechanism for row creation.
      Type Parameters:
      A - type of the target Address
      Parameters:
      target - the Target SNMP entity for the operation.
      rowStatusColumnOID - the column OID of the RowStatus column (without any instance identifier).
      rowIndex - the OID denoting the index of the table row to create.
      values - the values of columns to set in the row. If values is null the row is created via the tripple mode row creation mechanism (RowStatus is set to createAndWait). Otherwise, each variable binding has to contain the OID of the columnar object ID (without any instance identifier) and its value. On return, the variable bindings will be modified so that the variable binding OIDs will contain the instance OIDs of the respective columns (thus column OID + rowIndex).
      Returns:
      ResponseEvent the ResponseEvent instance returned by the SNMP session on response of the internally sent SET request. If null, an IO exception has occurred. Otherwise, if the response PDU is null a timeout has occurred. Otherwise, check the error status for SnmpConstants.SNMP_ERROR_SUCCESS to verify that the row creation was successful.
      Since:
      1.6

    • destroyRow

      public <A extends Address> ResponseEvent<A> destroyRow(Target<A> target, OID rowStatusColumnOID, OID rowIndex)
      Destroys an SNMP table row from a table that support the RowStatus mechanism for row creation/deletion.
      Type Parameters:
      A - address type of the target.
      Parameters:
      target - the Target SNMP entity for the operation.
      rowStatusColumnOID - the column OID of the RowStatus column (without any instance identifier).
      rowIndex - the OID denoting the index of the table row to destroy.
      Returns:
      ResponseEvent the ResponseEvent instance returned by the SNMP session on response of the internally sent SET request. If null, an IO exception has occurred. Otherwise, if the response PDU is null a timeout has occurred, Otherwise, check the error status for SnmpConstants.SNMP_ERROR_SUCCESS to verify that the row creation was successful.
      Since:
      1.7.6

    • getRowLimit

      public int getRowLimit()
      Gets the current row limit. A value greater than zero limits the total number of row events (TableEvent to the given number. See setRowLimit(int) for details.
      Returns:
      the row limit.
      Since:
      3.7.4

    • setRowLimit

      public void setRowLimit(int rowLimit)
      Sets the maximum number of rows returned from the target when getTable(Target, OID[], OID, OID) or any other overloaded variants of it are called. Please note, that the last rows returned up to getMaxNumRowsPerPDU() might not be complete (i.e., not all columns with data in the agent might have corresponding VariableBindings in the returned TableEvent. This can happen only for sparse tables, where not all columns of a row have values.
      Parameters:
      rowLimit - a value greater than zero limits the total number of rows returned. If the limit is reached and probably more rows would have been available the RetrievalEvent.STATUS_ROW_LIMIT_REACHED is returned.
      Since:
      3.7.4