1 karl 1.1.2.1 Using the CIM/XML Pull Operations
2
3 STATUS
4
|
5 karl 1.1.2.4 <<< The TODO section is being maintained during the review and checkin process
|
6 karl 1.1.2.1 to keep track of problems, errors, notes, etc. Must be deleted before
7 checkin to head of tree. Please feel free to add notes, etc in this
8 section as you review/test.>>>>>>
9
|
10 karl 1.1.2.4 TODO list:
|
11 karl 1.1.2.3 1. Binary operation from OOP. Need to add counter to binary
12 protocol to be able to count objects in response. Generates
13 warnings in things like messageserializer and does not work with
14 OOP right now.
|
15 karl 1.1.2.4 2. OpenExecQuery - Code is incorrect in that it used InstancesWithPath
16 where the spec is instances with no path. Need new function to wrap
17 getInstanceElement(withoutPathElement) in XmlReader. Note that
18 Alternate is to put flag on InstancesWith Path to say no path
19 3. Code for Pull part of OpenQueryInstancesRequest a) should be part of
20 the common CIMOperationRequestDispatcher execCommon code.
21 4. The changes to WQLCIMOperationRequestDispatcher and CQL... for handling
22 pull not completed so we feed the responses back to the EnmerationContext
23 queues
|
24 karl 1.1.2.7 3. Lots of minor TODOs, diagnostics, etc.
|
25 karl 1.1.2.12 4. External runtime variables. Proposing that they be fixed for this release.
|
26 karl 1.1.2.3 5. Decision on EnumerationContext timeout (separate thread or just
|
27 karl 1.1.2.4 checks during other operations). Can we, in fact really keep the
28 enumeration context table and queue under control without monitoring
29 with a separate thread. We must monitor for:
30 a. Client operation that stop requesting (i.e. inter operation time
31 exceeds operationTimeout). Note that if it simply exceeds the time
32 the next operation does the cleanup. The issue is those clients that
33 simply stop and do not either close or go to completion.
34 b. We should protect against providers that no not every finish delivering
35 or take to long between deliveries. This does not exist in Pegasus
36 today
|
37 karl 1.1.2.8 6. Clean up code in Dispatcher. Want to at least reduce the code for the
38 Open Operations to a set of templates so we know that the code is the
|
39 karl 1.1.2.12 same for all operations. Most of the common code for the enum, assoc,
40 ref operations and their corresponding open/pull operations is now in a
41 set of template functions so that it is largely common code. Used the
42 templates where there is common code but not the same CIMMessage. However,
43 we should be able to do some of it with common functions (i.e. not templates)
44 by adding a layer to CIMMessage hiearchy and reduce code size. The
|
45 karl 1.1.2.8 pull is a single template for both pullInstances and pullInstancePaths.
46 However, we might be able to reduce this to a single function by adding
47 a new level to CIMMessage.h (CommonPullResponse)
|
48 karl 1.1.2.3 7. Extension to avoid double move of objects in CIMResponseData (one
49 into enumerationContext queue and second to new cimResponseData for
50 response. Want to avoid second move by extending Open/Pull response
51 messages to include count and CIMResponse data to count objects out
52 of queue when converting (avoids the second move). Big issue here
53 with binary data since need to extend format to count it.
|
54 karl 1.1.2.8 8. NEXT TASKS:
55 a. get the pull operations rather than a template into a single
56 function by creating a new CIMPullResponse message in CIMMessage.h that
57 contains the pull data. Then we can use a single function to process all
58 pull operations.
|
59 karl 1.1.2.9 b. test the enumeration timeout thread
|
60 karl 1.1.2.11 d. Sort out issues with statisticalData since it really does not work
|
61 karl 1.1.2.12 with Open and Pull operations. See separate bug 9786.
62
63 30 September 2013 - CVS Update
64 Mergeout head of tree up to 29 September 2013.
|
65 karl 1.1.2.10
66 29 September 2013. CVS update.
67 1. Modified calls to statisticalData.cpp to a) directly call with request
68 type, b) incorporate the open, pull, etc. messages. However, since these
69 are not part of the CIM class, we must do something special with them.
|
70 karl 1.1.2.13 See bug 9785 for full solution to this issue.
|
71 karl 1.1.2.10 2. Corrected OOP interface to enable new flag to indicate internal operations
72 and set host, etc.
73 3. Add code to CQLOperationsDispatcher and WQLOperationDispatcher to clean
74 up CIMResponseDataCounter after filtering.
75 4. Modified ProviderAgent to set Host info for some pull operations.
76 5. Added new flag to CIMBinMsgSerializer and Deserializer.
|
77 karl 1.1.2.8
|
78 karl 1.1.2.9 17 September 2013 CVS update (Actually two different updates over 3 days)
79 1. Clean up some issues in CIMMessage.h and CIMMessage.cpp
80 2. Extend OpenExecQuery to WQL and CQL processors but return not complete
81 3. Remove memory leak in EnumerationContext and EnumerationContextTable
82 handling.
83 4. Created template functions for much of the pull operations.
84 5. Reversed order of queryLanguage and query (and changed names to match
85 execQuery) in client and server. Note that these are the execQuery
86 WQL and CQL filters and NOT FQL filters.
87 6. Some code cleanup in dispatcher
88 7. Today, passes all tests in pullop but issue in alltests. For some reason
89 not finding CIMObjectManager instance. Also, leaves enumeration contexts
90 if client terminates since cleanup thread not operating.
|
91 karl 1.1.2.11 8. XML from OOP not correctly processed.
|
92 karl 1.1.2.9
|
93 karl 1.1.2.8 14 September 2013 CVS update
94 Merged out up to 25 August. Cleaned up all operations and standardized code.
95 At this point the non pull operations code is in a set of templates but the
96 pull is not yet.
97 Fixed a significant number of problems so that it appears that the operations
98 except for OpenExecQuery run stably, at least with the pullop test program.
99 Note that there is a problem in that the Interop control provider is not
100 returning its singleton wbemserver object for some reason. Causes a test
101 failure
|
102 karl 1.1.2.7
103 Fixed for 16 June CVS Update
104 1. Cleaned up the enumerationContext and Table release functions and tested
105 to confirm that we do not lose memory in either normal sequences or
106 sequences that close early. Cleaned up pullop and added more tests
|
107 karl 1.1.2.8 Taged Before: PREAUG25UPDATE and after POSTAUG25UPDATE
|
108 karl 1.1.2.4
109 Fixed for 9 June CVS update
110 1. Cleaned up code for OpenQueryInstances. Note that this is incomplete.
111 No support in WQL or CQL Operations
112 2.
113
114 What was fixed for 5 June checkin.
|
115 karl 1.1.2.3 1. Extended ResponseTest MOF for for both CMPI and C++ subclasses
116 2. Fixed issues with pullop.
117 3. Fixed temp issue with CIMResponseData size by putting in mutex. That
118 is not a permanent fix but it gets around issue probably in the control
119 of the move logic that meant counts were off.
120 4. Fixed issues in Dispatcher so that associator code works. Still messy
121 code in the dispatcher.
122 5. Changed name of Enumerationtable.h & cpp to EnumerationContextTable.*
123 6 Changed name of ResponseStressTest module, classes, etc.
124
125 TAG: TASK_PEP317_5JUNE_2013_2
126
|
127 karl 1.1.2.2 2 June 2013
|
128 karl 1.1.2.1
129 Issues - KS
130 1. have not installed the binary move in CIMResponseData. Please run
131 with OPP off.
132 2. Some problem in the processing so we are getting server crashes.
133 Right no I am guessing that this is in the binaryCodec and am going to
134 expand the test tools to allow testing through the localhost.
135
136 3. Still way to many TODO and KS comments and KS_TEMPS. Removing bit by bit.
137
138 4. Env variable connection for the config parameters not installed.
139
140 5. Issue with the threaded timer. For some reason during tests it
141 eventually calls the timer thread with trash for the parm (which is
142 pointer to the EnumerationTable object). Caught because we do a valid
143 test at beginning of the function.
144
|
145 karl 1.1.2.2 6. Still using the templates in CIMOperationRequestDispatcher to simplify
146 the handle... processing.
147
148 7. I think I have a way around the double move of objects in the
149 EnumerationContext so that the outputter will just take a defined number
150 of objects directly from the gathering cache and save the second move.
151
152 8. Not yet passing all tests but getting closer now.
153
154 9. Created a tag before this commit TASK_PEP317_1JUNE_2013.
155
156 10. Next Tag will be TASK_PEP317_2_JUNE_2013 in the task branch
157
158
|
159 karl 1.1.2.1 ===========================================
160
161 OVERVIEW:
162
163 The operation extensions for pull operations defined in the DMTF specification
164 DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
165 including Client and Server.
166
167 These operations extend the CIM/XML individual operations to operation
168 sequences where the server must maintain state between operations in a
169 sequence and the client must execute multiple operations to get the full
170 set of instances or instance paths.
171
172 The following new CIM/XML operations as defined in DSP0200 are included;
173
174 -OpenEnumerateInstances
175 -openEnumerateInstancePaths
176 -OpenReferenceInstances
177 -OpenReferenceInstancePaths
178 -OpenAssociatiorInstances
179 -OpenAssociatorInstancePaths
180 karl 1.1.2.1 -PullInstancesWithPath
181 -PullInstancePaths
182 -CloseEnumeration
183 -EnumerationCount
|
184 karl 1.1.2.2 OpenExecQuery
|
185 karl 1.1.2.1
186 The following operations have not been implemented in this version of Pegasus:
187
188 -OpenQueryInstances
189
190 The following limitations on the implementation exist;
191
192 1. The filterQueryLanguage and filterQuery parameters are processed by
193 the Pegasus client but the server returns error if there is any data in
|
194 karl 1.1.2.2 either parameter. This work does not include the development of the
195 query language. Note that a separate effort to extend Pegasus to use
196 the DMTF FQL query language is in process.
|
197 karl 1.1.2.1
198 2. The input parameter continueOnError is processed correctly by the client
199 but the Pegasus server only provides for false since the server does not
200 include logic to continue processing responses after an error is
201 encountered.
202 This is consistent with the statement in the specification that use of
203 this functionality is optional and the fact that the DMTF agrees that all
204 of the issues of continuing after errors have not been clarified.
205
206 3. The operation enumerationCount is not processed by the server today since
207 a) really getting the count would be the same cost as the corresponding
208 enumeration, b) the server does not include a history or estimating
209 mechanism for this to date.
|
210 karl 1.1.2.2 NOTE: After a through review as part of the development of the next version
211 of CMPI we have concluded that this operation is probably not worth the
212 effort. Since it is optional, Pegasus will only return the unknown status
213 at this point
|
214 karl 1.1.2.1
215 Since the concept of sequences of operations linked together (open, pull, close)
216 is a major extension to the original CIM/XML operation concept of completely
217 independent operations several new pieces of functionality are implemented
218 to control interOperationTimeouts, counts of objects to be returned, etc.
219
220 TBD - Review this
221
222 CLIENT
223
224 The new operations follow the same pattern as the APIs for existing operations
225 in that:
226
227 1. All errors are handled as CIMException and Exception
228
229 2. The means of inputting parameters are the same except that there are
230 significantly more input parameters with the open operations and for the
231 first time operations return parameters as well as objects in the
232 response. Specifically the open and pull operations return values for
233 enumerationContext which is the identity for a pull sequence and
234 endOfSequence which is the marker the server sends in open and pull
235 karl 1.1.2.1 responses when it has no more objects to send.
236
237 The significant differences include:
238
239 1. Processing of parameters on responses (i.e. the endOfSequence and
240 enumerationContext parameters are returned for open and pull operations).
241
242 2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
243 cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
244 client api.
245
246 3. The association and reference operations ONLY process instances. They do
247 not include the capability to return classes like reference and associator
248 do and therefore return CIMInstance rather than CIMObject.
249
250 4. Paths are returned in all cases (i.e OpenEnumerateInstances and
251 PullInstancesWithPath where they were not with EnumeratInstances.
252
253 5. The client must maintain state between operations in a sequence (using
254 the enumerationContext parameter).
255
256 karl 1.1.2.1 TBD- Are there more differences.
257
258
259 SERVER
260
261 The Pegasus server attempts to always deliver the requested number of objects
262 for any open or pull request (the specification allows for the server to
263 deliver less than the requested number of objects and specifically to return
264 zero objects on open). We felt that it was worth any extra cost in processing
265 to provide the client with exactly what it had requested.
266
267 The pegasus server always closes an enumeration sequence upon receipt of any
268 error from the providers, repository, etc. Therefore the server will reject
|
269 karl 1.1.2.2 any request that has continueOnError = true;
270
271 Expansion to allow the continue on error may be added in a future version.
272 In any case, the whole purpose of the continue on error is really to allow
273 input from good providers to be mixed with providers that return errors so
274 that generally this would mean simply changing the logic in the return mechanism
275 to not shutdown when an error is received from any given provider.
276
277 Generally we do not believe that the providers need to do much more in the
278 future to support the continueOnError other than possibly allowing the provider
279 to continue processing after it has received an error.
|
280 karl 1.1.2.1
281 PROVIDERS
282
283 This implementation requires NO changes to the existing providers. The
284 provider APIs operate just as they do with the original operations.
285
286 Because the server processing is different however, there may be some
287 behavior differences primarily because the client now controls the speed of
288 delivery of objects.
289
290 In previous versions of Pegasus, the server attempts to deliver objects as
291 rapidly as then can be put on the network. In the case of HTTP chunked requests
292 they are delivered in chunks of about 100 objects. The primary delay for the
293 providers was the processing of each segment through the server. The server
294 is blocked so that no other segment can proceed through the server until that
295 segment is processed and sent on the network.
296 In the case of non-chunkedresponses, they are completely gathered in the serve
297 and then delivered as one non-chunked response. There were no delays for the
298 providers, just lots of possible memory use in the server.
299
300 The responses from providers (delivered through the deliver(...) interface are
301 karl 1.1.2.1 gathered into segments of about 100 objects and this group of objects is moved
302 through the server to be delivered to the client.
303
304 However with the inclusion of the pull operations, The segments of objects
305 from the providers are cached in the server response path until the
306 maxObjectCount for that request (open or pull) and that number returned in a
307 non-chunked response. Thus, if the client is slow to issue pull requests,
308 the providers might be delayed at some point to reduce memory usage in the
309 server (the delay appears as slow response tothe deliver operation).
310
311 In other words, the time to process large sets of responses from the provider
312 now depends on the speed of handling the client.
313
314 It is important to remember in developing providers that the Pegasus server
315 can most efficiently process responses if they are passed from the provider
316 to the server individually or in small arrays of objects rather than the
317 provider gathering very large arrays of objects and sending them to the
318 server.
319
|
320 karl 1.1.2.2 NEXT GENERATION PROVIDERS
321 KS_TODO
322
|
323 karl 1.1.2.1 CONFIGURATION PARAMETERS
324
325 The server includes several configuration parameters to set limits on the
326 processing of pull operations. All of these configuration parameters are
327 compile time parameters rather than runtime.
328
329 1. Maximum value of minimum interoperation time. This parameter defines the
330 maximum time allowed between the return of an open or pull response and
331 the receipt of the next pull or a close operation before the server may
332 close the enumeration. The specification allows the server to set a
333 maximum interoperation time and refuse open requests that with requested
334 operationTimeout greater than that time.
335 CIM_ERR_INVALID_OPERATION_TIMEOUT
336
337 This value is set with the Pegasus environment variable
338 PEGASUS_PULL....
339
340 2. Maximum objects returned in a single open or pull operation. The server
341 can set a maximum limit on the number of objects that can be returned in
342 a single open or pull oepration with the maxObjectCount parameter.
343
344 karl 1.1.2.1 3. Whether the server allows 0 as an interoperation timeout value. The value
345 zero is s special value for the interoperationTimeout in that it tells the
346 server to not timeout any enumeration sequence.
347
348 With this value for interoperationTimeout, the only way to close an
349 enumeration sequence is to complete all of the pulls or issue the close.
350 If for some reason the sequence is not completed, that enumeration context
351 would remain open indefinitly. Since in Pegasus any open enumeration
352 context uses resources (the context object and any provider resposnes that
353 have not yet been issued in a response) it would appear that most
354 platforms would not want to allow the existence of enumeration contexts
355 that cannot be closed by the server.
356
357 4, maximum consecutive pull requests with 0 maxObjectCount. The use of the
358 pull operation with maxObjectCount set to zero could be used to keep an
359 enumeration context open indefinitly (this tells the server to restart the
360 interoperationTimeout but not send any objects in the response). Therefore the
361 specification allows for the server setting maximum limits on this behavior
362 and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
363 exceeded.
364 Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
365 karl 1.1.2.1 a non-zero count resets this counter.
366
367 KS-TBD - Is this really logical since we can still block by just issuing
368 lots of zero request and an occansional request for one object.
369
370 Pegaus sets the value of this limit to 1000 and allows the implementer to
371 modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
372
373 5. Default operationTimeout -
374
375 The default of this parameter is to refuse operat
376
377 In the current release of Pegasus these are all compile time parameters.
|
378 karl 1.1.2.11
379
380 NOTES On working with task branch.
381
382 Merge out Process
383
384 To keep our TASK branch in sync with the current head of tree we need
385 to do a regular merge out. the TaskMakefile contains the makefile
386 procedures to do this efficiently. NOTE: Following these procedures is
387 important in that you are merging out new material each time you do
388 the merge out. If you were just to repeatedly merge out, you would be
389 merging previously merged changes a second time causing a real mess.
390
391 Start with new directory and put TaskMakefile above pegasus (needed so you
392 have this file for the initial operations.
393
394 make -f TaskMakefile branch_merge_out BNAME=PEP317-pullop ## takes a long time
395
396 This checks out current head, merges it into task branch and sets tags
397 for the mergeout. Note that at the end of this step this work is
398 part of the TASK... branch.
399 karl 1.1.2.11
400 NOW check for conflicts, errors, etc. that resulted from the merge.
401 Look for conflict flags, compare the results (I use linux merge as a
402 good graphic compare tool) and build and test. When you are satisfied
403 that the merge out is clean, you can commit the results to the TASK...
404 branch
405
406 To commit the work to this into Task branch
407
408 make -f mak/TaskMakefile branch_merge_out_commit BNAME=PEP317-pullop
409
410 or manually commit and finish as follows
411
412 cvs commit
413 make -f mak/TaskMakefile branch_merge_out_finish BNAME=PEP317-pullop
414
415 ## This last step is important since it cleans up temporary tags to prepare
416 you for the next checkout
417
418 COMPARE TASKBRANCH WITH HEAD
419
420 karl 1.1.2.11 In a new pegasus work space do same as above for merge out.
421
422 make -f TaskMakefile BNAME=PEP317-pullop
423
424 This produces a result which is all of the head merged into the branch.
425 A diff of this is all the new changes to the head of tree that you will
426 include into the merge.
427
|