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 karl 1.1.2.14 OOP right now. Corrected by converting to XML.
|
15 karl 1.1.2.15 2. OpenExecQuery - Code is incorrect in that it does not include the
16 return from the exec query function to the aggregator yet.
|
17 karl 1.1.2.4 3. Code for Pull part of OpenQueryInstancesRequest a) should be part of
18 the common CIMOperationRequestDispatcher execCommon code.
19 4. The changes to WQLCIMOperationRequestDispatcher and CQL... for handling
20 pull not completed so we feed the responses back to the EnmerationContext
21 queues
|
22 karl 1.1.2.18 3. Lots of minor TODOs, diagnostics, etc. still in the code
|
23 karl 1.1.2.14 4. External runtime variables. Proposing that they be fixed for this release
|
24 karl 1.1.2.15 rather than set by configuration. This should be discussed. Am making
|
25 karl 1.1.2.16 this a separate bug. See bug 9819 for the changes to cover this.
|
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 1.1.2.15 6. Consider moving some of the code in dispatcher from templates to common
|
38 karl 1.1.2.16 functions which would mean adding intermediate classes in CIMMessage but
39 would reduce code size.
|
40 karl 1.1.2.3 7. Extension to avoid double move of objects in CIMResponseData (one
41 into enumerationContext queue and second to new cimResponseData for
42 response. Want to avoid second move by extending Open/Pull response
43 messages to include count and CIMResponse data to count objects out
44 of queue when converting (avoids the second move). Big issue here
45 with binary data since need to extend format to count it.
|
46 karl 1.1.2.8 8. NEXT TASKS:
|
47 karl 1.1.2.15 a. test the enumeration timeout thread
|
48 karl 1.1.2.16 b. finish and test the OpenQueryInstances
49 c. Clean up TODOs
50 d. Find issue when we run makepoststarttests in pullop client with
51 forceProviderProcesses = true. This causes an operation like
52 cimcli pei CIM_ManagedElement to not complete (client timeout)
53 sometimes.
54
|
55 karl 1.1.2.21 17 April 2014
56 1. Reorganized the lock/unlock for handling the EnumerationContext because
57 there were thread windows (i.e. using the context after it was released
58 by other thread).
59 2. Modified code for the InteroperationTimerThread and deleted old code that
60 allowed this whole thread to by bypassed.
61 3. Internationalized a number of CIMOperationRequestDispatcher CIMException
62 response messages.
|
63 karl 1.1.2.22 4. Fixed stress tests in Client/tests/pullop that were getting timeout. The
64 issue was not the pull code but that fact that the idle client timeout
65 appears to be measureing to the end of the operation and the operation
66 was taking more than 40 seconds which is the default timeout. The idle
67 timeout takes no account of the fact that chunks are being regularly
68 delivered.
69 5. Several minor code changes.
70 NOTE: At this point the code passes extensive tests of the pull operations
71 both in OOP and none OOP mode (but with lots of diagnostics that may
72 be affecting timing).
|
73 karl 1.1.2.21
|
74 karl 1.1.2.20 10 April 2014
75 1. Fix issues causing failure with repeated pull operations.
76 2. Consolidate some of the CIMOperationRequestDispatcher.cpp template code
77
|
78 karl 1.1.2.19 7 April 2104
79 MergeOut of head of tree.
80
81 7 April 2014
82 1. Permanently fixed issue with getting correct number of objects on response.
83 This was introduced when did a temp fix to the OOP processing, in particular
84 removed the condition variable wait in the CIMOperationDispatcher handle
85 functions for opens and pulls. This moved the functionality to kick off
86 open and pull responses to the provider threads if there are not enough
87 objects when the handle is processed. Also added a variable to allow us to
88 test with either a) responses are required to satisfy the original
89 request size or) b, responses are required only to return some objects.
90 We will add statistics to see which of these works best.
91 2. Modified WsmProcessor EnumerationContext class since the name conflicts
92 with the pull Operation EnumerationContext class.
93 3. Created a new intermediate level of CIMMessage, the CIMPullResponseData
94 Message so that all of the open and Pull responses can use common code
95 since there are only two variables (endOfSequence and EnumerationContext)
96 and they are common across all the open and pull responses.
97 4. Removed a number of diagnostics.
98 5. Cleaned up the Dispcatcher so that the open and pulls have common response
99 karl 1.1.2.19 code and that code can be used from the dispatcher handle functions and
100 the provider response functions.
101
|
102 karl 1.1.2.18 31 March 2014 - Checkin
103 1. Fixed issues in OOP processing of pull operations, in particular
104 issues with cimxml output format when processed through the
105 *InternalXmlEncoder functions.
106 2. Clean up some of the internalXml functionality
107 3. Found issues causing timeout with a particular provider. The issue
108 is that the dispatcher and monitor end up using the same thread so the
109 condition variable in the dispatcher thread stops the monitor. Turned
110 off the conditionVariable in getCache for the moment which means that
111 we get number of responses for open... with 0 objects before the
112 providers can begin to respond. This is only for test.
113 4. Added some statistics for enumerations and display the statistics
114 when we close the server (same as cache statistics)
115
116 12 March 2014 - Mergeout and Mergein
117 1. Mergeout to head of tree for this date and mergein for patch update
118 to bug 9676
119 2. Extensions to pullop tests program and tests.
120 3. Added some diagnostics in looking for OOP issue.
121 4. Removed a number of diagnostics messages and cleaned up code in
122 dispatcher to simplify pull operation processing.
123 karl 1.1.2.18
|
124 karl 1.1.2.17 15 December 2013
125 1. Mergeout and mergein up to 15 December 2013
126 2. Clean up issues from tests documented in bug 9676 last week.
127 3. Clean up some code in dispatcher
128 4. Remove the filter function from ResponseStressc++Provider.
129
|
130 karl 1.1.2.16 21 November 2013
131 1. Mergeout from head of tree to 21 November 2013.
|
132 karl 1.1.2.15
133 18 November 2013
134 1. Cleanup of a bunch of minor errors and completion of all of the code for
135 the openQueryInstances except for the PullInstances in Dispatcher and
136 the aggregator function.
137 2. OpenqueryInstances added to cimcli.
|
138 karl 1.1.2.14
139 13 October 2013 CVS branch update.
140 1. Integrated bug 9786 into the branch. Note that we need to test the
141 generated statistics.
142 2. Mergeout executed to update to head of tree as of 8:00 am 13 October 2013.
143 3. Cleaned up several errors in OOP processing. Note that there is at least
144 one issue left when we to a pull on ManagedElement in at least one of the
145 namespaces.
146 4. Cleaned up some of the outstanding diagnostic code
147 5. Generally passes all tests except for one test of pullop where it is trying
148 to pull enum instances CIM_ManagedElement from a particular namespace.
|
149 karl 1.1.2.12
|
150 karl 1.1.2.16 NOTE: I did not make comments here for changes in October despite the fact
151 that I did 2 mergouts, number of fixes, and a mergein.
152
|
153 karl 1.1.2.12 30 September 2013 - CVS Update
154 Mergeout head of tree up to 29 September 2013.
|
155 karl 1.1.2.10
156 29 September 2013. CVS update.
157 1. Modified calls to statisticalData.cpp to a) directly call with request
158 type, b) incorporate the open, pull, etc. messages. However, since these
159 are not part of the CIM class, we must do something special with them.
|
160 karl 1.1.2.13 See bug 9785 for full solution to this issue.
|
161 karl 1.1.2.10 2. Corrected OOP interface to enable new flag to indicate internal operations
162 and set host, etc.
163 3. Add code to CQLOperationsDispatcher and WQLOperationDispatcher to clean
164 up CIMResponseDataCounter after filtering.
165 4. Modified ProviderAgent to set Host info for some pull operations.
166 5. Added new flag to CIMBinMsgSerializer and Deserializer.
|
167 karl 1.1.2.8
|
168 karl 1.1.2.9 17 September 2013 CVS update (Actually two different updates over 3 days)
169 1. Clean up some issues in CIMMessage.h and CIMMessage.cpp
170 2. Extend OpenExecQuery to WQL and CQL processors but return not complete
171 3. Remove memory leak in EnumerationContext and EnumerationContextTable
172 handling.
173 4. Created template functions for much of the pull operations.
174 5. Reversed order of queryLanguage and query (and changed names to match
175 execQuery) in client and server. Note that these are the execQuery
176 WQL and CQL filters and NOT FQL filters.
177 6. Some code cleanup in dispatcher
178 7. Today, passes all tests in pullop but issue in alltests. For some reason
179 not finding CIMObjectManager instance. Also, leaves enumeration contexts
180 if client terminates since cleanup thread not operating.
|
181 karl 1.1.2.11 8. XML from OOP not correctly processed.
|
182 karl 1.1.2.9
|
183 karl 1.1.2.8 14 September 2013 CVS update
184 Merged out up to 25 August. Cleaned up all operations and standardized code.
185 At this point the non pull operations code is in a set of templates but the
186 pull is not yet.
187 Fixed a significant number of problems so that it appears that the operations
188 except for OpenExecQuery run stably, at least with the pullop test program.
189 Note that there is a problem in that the Interop control provider is not
190 returning its singleton wbemserver object for some reason. Causes a test
191 failure
|
192 karl 1.1.2.7
193 Fixed for 16 June CVS Update
194 1. Cleaned up the enumerationContext and Table release functions and tested
195 to confirm that we do not lose memory in either normal sequences or
196 sequences that close early. Cleaned up pullop and added more tests
|
197 karl 1.1.2.8 Taged Before: PREAUG25UPDATE and after POSTAUG25UPDATE
|
198 karl 1.1.2.4
199 Fixed for 9 June CVS update
200 1. Cleaned up code for OpenQueryInstances. Note that this is incomplete.
201 No support in WQL or CQL Operations
202 2.
203
204 What was fixed for 5 June checkin.
|
205 karl 1.1.2.3 1. Extended ResponseTest MOF for for both CMPI and C++ subclasses
206 2. Fixed issues with pullop.
207 3. Fixed temp issue with CIMResponseData size by putting in mutex. That
208 is not a permanent fix but it gets around issue probably in the control
209 of the move logic that meant counts were off.
210 4. Fixed issues in Dispatcher so that associator code works. Still messy
211 code in the dispatcher.
212 5. Changed name of Enumerationtable.h & cpp to EnumerationContextTable.*
213 6 Changed name of ResponseStressTest module, classes, etc.
214
215 TAG: TASK_PEP317_5JUNE_2013_2
216
|
217 karl 1.1.2.2 2 June 2013
|
218 karl 1.1.2.1
219 Issues - KS
220
|
221 karl 1.1.2.15 - Still way to many TODO and KS comments and KS_TEMPS. Removing bit by bit.
|
222 karl 1.1.2.1
|
223 karl 1.1.2.15 - Runtime variable connection for the config parameters not installed. That
224 has been made into a separate bug (see bug 9819)
|
225 karl 1.1.2.1
226 5. Issue with the threaded timer. For some reason during tests it
227 eventually calls the timer thread with trash for the parm (which is
228 pointer to the EnumerationTable object). Caught because we do a valid
229 test at beginning of the function.
230
|
231 karl 1.1.2.2 6. Still using the templates in CIMOperationRequestDispatcher to simplify
232 the handle... processing.
233
234 7. I think I have a way around the double move of objects in the
235 EnumerationContext so that the outputter will just take a defined number
236 of objects directly from the gathering cache and save the second move.
237
|
238 karl 1.1.2.15 8. Not yet passing all tests but getting closer now. The major test that is
239 causing an error today is the execution of a full enumeration with the
240 forceProviders = true. This causes a client timeout sometimes.
|
241 karl 1.1.2.2
242
243
|
244 karl 1.1.2.1 ===========================================
245
246 OVERVIEW:
247
248 The operation extensions for pull operations defined in the DMTF specification
249 DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
250 including Client and Server.
251
252 These operations extend the CIM/XML individual operations to operation
253 sequences where the server must maintain state between operations in a
254 sequence and the client must execute multiple operations to get the full
255 set of instances or instance paths.
256
257 The following new CIM/XML operations as defined in DSP0200 are included;
258
259 -OpenEnumerateInstances
260 -openEnumerateInstancePaths
261 -OpenReferenceInstances
262 -OpenReferenceInstancePaths
263 -OpenAssociatiorInstances
264 -OpenAssociatorInstancePaths
|
265 karl 1.1.2.16 -OpenQueryInstances
|
266 karl 1.1.2.1 -PullInstancesWithPath
267 -PullInstancePaths
|
268 karl 1.1.2.16 -PullInstances
|
269 karl 1.1.2.1 -CloseEnumeration
270 -EnumerationCount
|
271 karl 1.1.2.2 OpenExecQuery
|
272 karl 1.1.2.1
273 The following operations have not been implemented in this version of Pegasus:
274
275 -OpenQueryInstances
276
277 The following limitations on the implementation exist;
278
279 1. The filterQueryLanguage and filterQuery parameters are processed by
280 the Pegasus client but the server returns error if there is any data in
|
281 karl 1.1.2.2 either parameter. This work does not include the development of the
282 query language. Note that a separate effort to extend Pegasus to use
283 the DMTF FQL query language is in process.
|
284 karl 1.1.2.1
285 2. The input parameter continueOnError is processed correctly by the client
286 but the Pegasus server only provides for false since the server does not
287 include logic to continue processing responses after an error is
288 encountered.
289 This is consistent with the statement in the specification that use of
290 this functionality is optional and the fact that the DMTF agrees that all
291 of the issues of continuing after errors have not been clarified.
292
293 3. The operation enumerationCount is not processed by the server today since
294 a) really getting the count would be the same cost as the corresponding
295 enumeration, b) the server does not include a history or estimating
296 mechanism for this to date.
|
297 karl 1.1.2.2 NOTE: After a through review as part of the development of the next version
298 of CMPI we have concluded that this operation is probably not worth the
299 effort. Since it is optional, Pegasus will only return the unknown status
300 at this point
|
301 karl 1.1.2.1
302 Since the concept of sequences of operations linked together (open, pull, close)
303 is a major extension to the original CIM/XML operation concept of completely
304 independent operations several new pieces of functionality are implemented
305 to control interOperationTimeouts, counts of objects to be returned, etc.
306
307 TBD - Review this
308
309 CLIENT
310
311 The new operations follow the same pattern as the APIs for existing operations
312 in that:
313
314 1. All errors are handled as CIMException and Exception
315
316 2. The means of inputting parameters are the same except that there are
317 significantly more input parameters with the open operations and for the
318 first time operations return parameters as well as objects in the
319 response. Specifically the open and pull operations return values for
320 enumerationContext which is the identity for a pull sequence and
321 endOfSequence which is the marker the server sends in open and pull
322 karl 1.1.2.1 responses when it has no more objects to send.
323
324 The significant differences include:
325
326 1. Processing of parameters on responses (i.e. the endOfSequence and
327 enumerationContext parameters are returned for open and pull operations).
328
329 2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
330 cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
331 client api.
332
333 3. The association and reference operations ONLY process instances. They do
334 not include the capability to return classes like reference and associator
335 do and therefore return CIMInstance rather than CIMObject.
336
337 4. Paths are returned in all cases (i.e OpenEnumerateInstances and
338 PullInstancesWithPath where they were not with EnumeratInstances.
339
340 5. The client must maintain state between operations in a sequence (using
341 the enumerationContext parameter).
342
343 karl 1.1.2.1 TBD- Are there more differences.
344
345
346 SERVER
347
348 The Pegasus server attempts to always deliver the requested number of objects
349 for any open or pull request (the specification allows for the server to
350 deliver less than the requested number of objects and specifically to return
351 zero objects on open). We felt that it was worth any extra cost in processing
352 to provide the client with exactly what it had requested.
353
354 The pegasus server always closes an enumeration sequence upon receipt of any
355 error from the providers, repository, etc. Therefore the server will reject
|
356 karl 1.1.2.2 any request that has continueOnError = true;
357
358 Expansion to allow the continue on error may be added in a future version.
359 In any case, the whole purpose of the continue on error is really to allow
360 input from good providers to be mixed with providers that return errors so
361 that generally this would mean simply changing the logic in the return mechanism
362 to not shutdown when an error is received from any given provider.
363
364 Generally we do not believe that the providers need to do much more in the
365 future to support the continueOnError other than possibly allowing the provider
366 to continue processing after it has received an error.
|
367 karl 1.1.2.1
368 PROVIDERS
369
370 This implementation requires NO changes to the existing providers. The
371 provider APIs operate just as they do with the original operations.
372
373 Because the server processing is different however, there may be some
374 behavior differences primarily because the client now controls the speed of
375 delivery of objects.
376
377 In previous versions of Pegasus, the server attempts to deliver objects as
378 rapidly as then can be put on the network. In the case of HTTP chunked requests
379 they are delivered in chunks of about 100 objects. The primary delay for the
380 providers was the processing of each segment through the server. The server
381 is blocked so that no other segment can proceed through the server until that
382 segment is processed and sent on the network.
383 In the case of non-chunkedresponses, they are completely gathered in the serve
384 and then delivered as one non-chunked response. There were no delays for the
385 providers, just lots of possible memory use in the server.
386
387 The responses from providers (delivered through the deliver(...) interface are
388 karl 1.1.2.1 gathered into segments of about 100 objects and this group of objects is moved
389 through the server to be delivered to the client.
390
391 However with the inclusion of the pull operations, The segments of objects
392 from the providers are cached in the server response path until the
393 maxObjectCount for that request (open or pull) and that number returned in a
394 non-chunked response. Thus, if the client is slow to issue pull requests,
395 the providers might be delayed at some point to reduce memory usage in the
396 server (the delay appears as slow response tothe deliver operation).
397
398 In other words, the time to process large sets of responses from the provider
399 now depends on the speed of handling the client.
400
401 It is important to remember in developing providers that the Pegasus server
402 can most efficiently process responses if they are passed from the provider
403 to the server individually or in small arrays of objects rather than the
404 provider gathering very large arrays of objects and sending them to the
405 server.
406
|
407 karl 1.1.2.2 NEXT GENERATION PROVIDERS
408 KS_TODO
409
|
410 karl 1.1.2.1 CONFIGURATION PARAMETERS
411
412 The server includes several configuration parameters to set limits on the
413 processing of pull operations. All of these configuration parameters are
414 compile time parameters rather than runtime.
415
416 1. Maximum value of minimum interoperation time. This parameter defines the
417 maximum time allowed between the return of an open or pull response and
418 the receipt of the next pull or a close operation before the server may
419 close the enumeration. The specification allows the server to set a
420 maximum interoperation time and refuse open requests that with requested
421 operationTimeout greater than that time.
422 CIM_ERR_INVALID_OPERATION_TIMEOUT
423
424 This value is set with the Pegasus environment variable
425 PEGASUS_PULL....
426
427 2. Maximum objects returned in a single open or pull operation. The server
428 can set a maximum limit on the number of objects that can be returned in
429 a single open or pull oepration with the maxObjectCount parameter.
430
431 karl 1.1.2.1 3. Whether the server allows 0 as an interoperation timeout value. The value
432 zero is s special value for the interoperationTimeout in that it tells the
433 server to not timeout any enumeration sequence.
434
435 With this value for interoperationTimeout, the only way to close an
436 enumeration sequence is to complete all of the pulls or issue the close.
437 If for some reason the sequence is not completed, that enumeration context
438 would remain open indefinitly. Since in Pegasus any open enumeration
439 context uses resources (the context object and any provider resposnes that
440 have not yet been issued in a response) it would appear that most
441 platforms would not want to allow the existence of enumeration contexts
442 that cannot be closed by the server.
443
444 4, maximum consecutive pull requests with 0 maxObjectCount. The use of the
445 pull operation with maxObjectCount set to zero could be used to keep an
446 enumeration context open indefinitly (this tells the server to restart the
447 interoperationTimeout but not send any objects in the response). Therefore the
448 specification allows for the server setting maximum limits on this behavior
449 and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
450 exceeded.
451 Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
452 karl 1.1.2.1 a non-zero count resets this counter.
453
454 KS-TBD - Is this really logical since we can still block by just issuing
455 lots of zero request and an occansional request for one object.
456
457 Pegaus sets the value of this limit to 1000 and allows the implementer to
458 modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
459
460 5. Default operationTimeout -
461
462 The default of this parameter is to refuse operat
463
464 In the current release of Pegasus these are all compile time parameters.
|
465 karl 1.1.2.11
466
467 NOTES On working with task branch.
468
469 Merge out Process
470
471 To keep our TASK branch in sync with the current head of tree we need
472 to do a regular merge out. the TaskMakefile contains the makefile
473 procedures to do this efficiently. NOTE: Following these procedures is
474 important in that you are merging out new material each time you do
475 the merge out. If you were just to repeatedly merge out, you would be
476 merging previously merged changes a second time causing a real mess.
477
478 Start with new directory and put TaskMakefile above pegasus (needed so you
479 have this file for the initial operations.
480
481 make -f TaskMakefile branch_merge_out BNAME=PEP317-pullop ## takes a long time
482
483 This checks out current head, merges it into task branch and sets tags
484 for the mergeout. Note that at the end of this step this work is
485 part of the TASK... branch.
486 karl 1.1.2.11
487 NOW check for conflicts, errors, etc. that resulted from the merge.
488 Look for conflict flags, compare the results (I use linux merge as a
489 good graphic compare tool) and build and test. When you are satisfied
490 that the merge out is clean, you can commit the results to the TASK...
491 branch
492
493 To commit the work to this into Task branch
494
495 make -f mak/TaskMakefile branch_merge_out_commit BNAME=PEP317-pullop
496
497 or manually commit and finish as follows
498
499 cvs commit
500 make -f mak/TaskMakefile branch_merge_out_finish BNAME=PEP317-pullop
501
502 ## This last step is important since it cleans up temporary tags to prepare
503 you for the next checkout
504
505 COMPARE TASKBRANCH WITH HEAD
506
507 karl 1.1.2.11 In a new pegasus work space do same as above for merge out.
508
509 make -f TaskMakefile BNAME=PEP317-pullop
510
511 This produces a result which is all of the head merged into the branch.
512 A diff of this is all the new changes to the head of tree that you will
513 include into the merge.
514
|