(file) Return to readme.pulloperations CVS log (file) (dir) Up to [Pegasus] / pegasus

Diff for /pegasus/readme.pulloperations between version 1.1 and 1.1.2.1

version 1.1, 2011/01/16 21:11:30 version 1.1.2.1, 2011/01/16 21:11:30
Line 0 
Line 1 
           Using the CIM/XML Pull Operations
   
   STATUS
   
   <<< This section is being maintained during the review and checkin process
   to keep track of problems, errors, notes, etc.  Must be deleted before
   checkin to head of tree. Please feel free to add notes, etc in this
   section as you review/test.>>>>>>
   
   16 Jan 2011
   
   Issues  - KS
   1. have not installed the binary move in CIMResponseData. Please run
   with OPP off.
   2. Some problem in the processing so we are getting server crashes.
   Right no I am guessing that this is in the binaryCodec and am going to
   expand the test tools to allow testing through the localhost.
   
   3. Still way to many TODO and KS comments and KS_TEMPS.  Removing bit by bit.
   
   4. Env variable connection for the config parameters not installed.
   
   5. Issue with the threaded timer.  For some reason during tests it
   eventually calls the timer thread with trash for the parm (which is
   pointer to the EnumerationTable object). Caught because we do a valid
   test at beginning of the function.
   
   ===========================================
   
   OVERVIEW:
   
   The operation extensions for pull operations defined in the DMTF specification
   DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
   including Client and Server.
   
   These operations extend the CIM/XML  individual operations to operation
   sequences where the server must maintain state between operations in a
   sequence and the client must execute multiple operations to get the full
   set of instances or instance paths.
   
   The following new CIM/XML operations as defined in DSP0200 are included;
   
       -OpenEnumerateInstances
       -openEnumerateInstancePaths
       -OpenReferenceInstances
       -OpenReferenceInstancePaths
       -OpenAssociatiorInstances
       -OpenAssociatorInstancePaths
       -PullInstancesWithPath
       -PullInstancePaths
       -CloseEnumeration
       -EnumerationCount
   
   The following  operations have not been implemented in this version of Pegasus:
   
       -OpenQueryInstances
   
   The following limitations on the implementation exist;
   
   1. The filterQueryLanguage and filterQuery parameters are processed by
      the Pegasus client but the server returns error if there is any data in
      either parameter.  This is because the definition of query language has not
      been clarified to date by the DMTF.
   
   2. The input parameter continueOnError is processed correctly by the client
      but the Pegasus server only provides for false since the server does not
      include logic to continue processing responses after an error is
      encountered.
      This is consistent with the statement in the specification that use of
      this functionality is optional and the fact that the DMTF agrees that all
      of the issues of continuing after errors have not been clarified.
   
   3. The operation enumerationCount is not processed by the server today since
      a) really getting the count would be the same cost as the corresponding
      enumeration, b) the server does not include a history or estimating
      mechanism for this to date.
      KS_TBD_ confirm this
   
   Since the concept of sequences of operations linked together (open, pull, close)
   is a major extension to the original CIM/XML operation concept of completely
   independent operations several new pieces of functionality are implemented
   to control interOperationTimeouts, counts of objects to be returned, etc.
   
   TBD - Review this
   
   CLIENT
   
   The new operations follow the same pattern as the APIs for existing operations
   in that:
   
   1. All errors are handled as CIMException and Exception
   
   2. The means of inputting parameters are the same except that there are
      significantly more input parameters with the open operations and for the
      first time operations return parameters as well as objects in the
      response.  Specifically the open and pull operations return values for
      enumerationContext which is the identity for a pull sequence and
      endOfSequence which is the marker the server sends in open and pull
      responses when it has no more objects to send.
   
   The significant differences include:
   
   1. Processing of parameters on responses (i.e. the endOfSequence and
      enumerationContext parameters are returned for open and pull operations).
   
   2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
      cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
      client api.
   
   3. The association and reference operations ONLY process instances.  They do
      not include the capability to return classes like reference and associator
      do and therefore return CIMInstance rather than CIMObject.
   
   4. Paths are returned in all cases (i.e OpenEnumerateInstances and
      PullInstancesWithPath where they were not with EnumeratInstances.
   
   5. The client must maintain state between operations in a sequence (using
      the enumerationContext parameter).
   
   TBD- Are there more differences.
   
   
   SERVER
   
   The Pegasus server attempts to always deliver the requested number of objects
   for any open or pull request (the specification allows for the server to
   deliver less than the requested number of objects and specifically to return
   zero objects on open).  We felt that it was worth any extra cost in processing
   to provide the client with exactly what it had requested.
   
   continueOnError Parameter
   The pegasus server always closes an enumeration sequence upon receipt of any
   error from the providers, repository, etc. Therefore the server will reject
   and request that has continueOnError = true;
   
   PROVIDERS
   
   This implementation requires NO changes to the existing providers.  The
   provider APIs operate just as they do with the original operations.
   
   Because the server processing is different however, there may be some
   behavior differences primarily because the client now controls the speed of
   delivery of objects.
   
   In previous versions of Pegasus, the server attempts to deliver objects as
   rapidly as then can be put on the network.  In the case of HTTP chunked requests
   they are delivered in chunks of about 100 objects. The primary delay for the
   providers was the processing of each segment through the server.  The server
   is blocked so that no other segment can proceed through the server until that
   segment is processed and sent on the network.
   In the case of non-chunkedresponses, they are completely gathered in the serve
   and then delivered as one non-chunked response. There were no delays for the
   providers, just lots of possible memory use in the server.
   
   The responses from providers (delivered through the deliver(...) interface are
   gathered into segments of about 100 objects and this group of objects is moved
   through the server to be delivered to the client.
   
   However with the inclusion of the pull operations,   The segments of objects
   from the providers are cached in the server response path until the
   maxObjectCount for that request (open or pull) and that number returned in a
   non-chunked response. Thus, if the client is slow to issue pull requests,
   the providers might be delayed at some point to reduce memory usage in the
   server (the delay appears as slow response tothe deliver operation).
   
   In other words, the time to process large sets of responses from the provider
   now depends on the speed of handling the client.
   
   It is important to remember in developing providers that the Pegasus server
   can most efficiently process responses if they are passed from the provider
   to the server individually or in small arrays of objects rather than the
   provider gathering very large arrays of objects and sending them to the
   server.
   
   CONFIGURATION PARAMETERS
   
   The server includes several configuration parameters to set limits on the
   processing of pull operations.  All of these configuration parameters are
   compile time parameters rather than runtime.
   
   1. Maximum value of minimum interoperation time.  This parameter defines the
   maximum time allowed between the return of an open or pull response and
   the receipt of the next pull or a close operation before the server may
   close the enumeration.  The specification allows the server to set a
   maximum interoperation time and refuse open requests that with requested
   operationTimeout greater than that time.
   CIM_ERR_INVALID_OPERATION_TIMEOUT
   
   This value is set with the Pegasus environment variable
   PEGASUS_PULL....
   
   2. Maximum objects returned in a single open or pull operation.  The server
   can set a maximum limit on the number of objects that can be returned in
   a single open or pull oepration with the maxObjectCount parameter.
   
   3. Whether the server allows 0 as an interoperation timeout value. The value
   zero is s special value for the interoperationTimeout in that it tells the
   server to not timeout any enumeration sequence.
   
   With this value for interoperationTimeout, the only way to close an
   enumeration sequence is to complete all of the pulls or issue the close.
   If for some reason the sequence is not completed, that enumeration context
   would remain open indefinitly.  Since in Pegasus any open enumeration
   context uses resources (the context object and any provider resposnes that
   have not yet been issued in a response) it would appear that most
   platforms would not want to allow the existence of enumeration contexts
   that cannot be closed by the server.
   
   4, maximum consecutive pull requests with 0 maxObjectCount.  The use of the
   pull operation with maxObjectCount set to zero could be used to keep an
   enumeration context open indefinitly (this tells the server to restart the
   interoperationTimeout but not send any objects in the response). Therefore the
   specification allows for the server setting maximum limits on this behavior
   and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
   exceeded.
   Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
   a non-zero count resets this counter.
   
   KS-TBD - Is this really logical since we can still block by just issuing
   lots of zero request and an occansional request for one object.
   
   Pegaus sets the value of this limit to 1000 and allows the implementer to
   modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
   
   5. Default operationTimeout -
   
   The default of this parameter is to refuse operat
   
   In the current release of Pegasus these are all compile time parameters.


Legend:
Removed from v.1.1  
changed lines
  Added in v.1.1.2.1

No CVS admin address has been configured
Powered by
ViewCVS 0.9.2