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 1.1.2.31 OOP right now. Fixed by converting to XML. Concluded that we do not
15 need to do this. The binary response is not really used often
16 in the current environment So double mapping it is not a major issue.
17 Leave this as FUTURE
|
18 karl 1.1.2.35 2. DONE OpenExecQuery - Code is incomplete in that it does not include the
|
19 karl 1.1.2.31 return from the exec query function to the aggregator yet. Code for Pull
|
20 karl 1.1.2.34 part of OpenQueryInstancesRequest a) should be part of the common
21 CIMOperationRequestDispatcher execCommon code. DONE, not tested. 29 May
|
22 karl 1.1.2.35 30 May tested with simple cli calls.
23 3. DONE The changes to WQLCIMOperationRequestDispatcher and CQL... for handling
|
24 karl 1.1.2.4 pull not completed so we feed the responses back to the EnmerationContext
|
25 karl 1.1.2.35 queues. DONE Not tested 29 May 30 May DONE, some testing done
26 4. Minor TODOs, diagnostics, etc. still in the code. Reduced almost to
27 none now. We are leaving some in as PEGASUS_DEBUG
|
28 karl 1.1.2.31 5. Extension to avoid double move of objects in CIMResponseData (one
|
29 karl 1.1.2.3 into enumerationContext queue and second to new cimResponseData for
30 response. Want to avoid second move by extending Open/Pull response
31 messages to include count and CIMResponse data to count objects out
|
32 karl 1.1.2.23 of queue when converting (avoids the second move). This would mean
33 extending the output writers to allow the count field to be supplied
|
34 karl 1.1.2.30 so they would only create output for up to the count supplied.(Make
|
35 karl 1.1.2.35 this future beyond bug 9676). This is fairly extensive because it
36 extends beyond CIMResponseData to SCMO and XML writers where the
37 XmlWriters used by encodeXmlResponse would have to have counters
38 added. Then instead of copying on getCache we would simply pass the
39 cache and count on and the writer would take and remove.
|
40 karl 1.1.2.31 6. Add more static tests (currently only OpenEnumerateInstances and
|
41 karl 1.1.2.30 OpenEnumerateInstanceNames covered).
|
42 karl 1.1.2.26 Low priority because most of this except exact XML covered in other
|
43 karl 1.1.2.28 tests. note that we can only add very limited tests since a) no
|
44 karl 1.1.2.31 way to test multiple operations in sequence(i.e. open/pull) and
|
45 karl 1.1.2.28 cannot be sure how many responses will come back in all cases even
|
46 karl 1.1.2.35 for open. The only reason for most of these tests is to assure that
47 we do not change XML inadvertendtly in future with change that is
48 compatible to both our server and client but breaks other clients or
49 servers. In any case we can do only minimal tests because most real
50 testing involves both open and pull and the static tests are by their
51 nature single operation.
52 7. Clean up the CIMOperationRequestDecoder.cpp reject CIMException
53 further since there is some overlap. The only overlap is that a
54 couple of the specific functions duplicate. VERY LOW priority.
|
55 karl 1.1.2.32 8. Correct issue between operations and HTTP where we are sending
|
56 karl 1.1.2.31 trailers with exceptions. Modify response mechanisms so that we
57 set non-chunked for all responses where we send error responses to
|
58 karl 1.1.2.35 avoid the trailers. NOTE: There should be now a bug on this in general
59 where we would want to send an initial error without the trailer. Should
60 have always done that.
|
61 karl 1.1.2.32 9. It would be more efficient in dispatcher to always use exception for
|
62 karl 1.1.2.31 rejects and change the _reject functions so that they never return
63 when they reject. This is VERY LOW PRIORITY and primarily saves
64 a few lines of code in the reject functions and their calls. Means we
65 would code.
|
66 karl 1.1.2.35 _rejectIfEnumerationToBroad(...);
67 _rejectIfThisParameterIncorrect(...);
|
68 karl 1.1.2.31
69 instead of
70 if (_rejectIfEnum...)
71 {
72 return true
73 }
74 It would mean that the method trace for the handlers would not return
|
75 karl 1.1.2.35 an exit if we rejected. VERY LOW PRIORITY. Possibly FUTURE. No behavior
76 change, just more compact source code but it messes with the method
77 trace logic.
|
78 karl 1.1.2.34 10. There are still a couple of template functions around the task of
|
79 karl 1.1.2.32 distributing requests to the multiple providers.
80 In fact there are two similiar but different templates for
|
81 karl 1.1.2.31 the associators(i.e. assoc, ref, etc.) functions and the parallel
82 openAssoc functions. It would be nice to consolidate that logic and
83 further to try to create a non-template form for those functions. LOW
84 PRIORITY.
|
85 karl 1.1.2.34 11. External runtime variables. This bug leaves these variables as
86 internal to CIMOperationRequestDispatcher.cpp.
87 See bug 9819 for the changes to externalize these variables, both
88 the runtime and buildtime variables. The Work group discussion
89 documents what the variables are and how they tie to statistics info.
90 See the OpenPegasus Wiki, Pull operations work group.
|
91 karl 1.1.2.35 12. Trace functions in CIMResponseData,h & cpp should be PEGASUS_DEBUG.
92 Same for trace function in EnumerationContext and EnumerationContextTable
93 13. Question. We added trace in CIMRequestOperationDispatcher.cpp if
94 query enum is recevied with no object path. However, since this is
95 a provider provblem should we be doing something other than a trace
96 (ex. log) since traces are often not really used and they do generate
97 lots of data whereby something like this could get lost. Also, need
98 to review what level of trace if we keep trace.
|
99 karl 1.1.2.34
|
100 karl 1.1.2.26
101 NEXT TASKS:
|
102 karl 1.1.2.35 a. finish 8, 12 above ( Finish by about 1 June). Do next mergeout and
103 mergein to create a complete patch
104
105 30 May 2014
106 1. Correct issue in UintArgs. Modified to use smart pointers, consistent
107 with public interfaces. Tested
108 2. Complete code for OpenQueryInstances and PullInstances so that the complete
109 sequence works for both the server and client (i.e. 2, 3 above done
110 except for more tests in pullop).
111 3. Fix issue in cimcli where the OpenQueryInstances was outputting a path
112 component where there is no path transmitted from the server.
113 4. Moved large block of common code in CQLOperationRequestDispatcher.cpp
114 and WQLOperationRequestDispatcher.cpp to common function
115 handleQueryRequestCommon in CIMOperationRequestDispatcher.cpp. These
116 handlers now reduced to just the parser processing and minimal error
117 checks.
118 5. Cleaned up some code in EnumerationContextTable.CreateContext so only
119 single interface for all requests.
120 6, Commented out a number of PEG_TRACE messages. Note that they are
121 commented with 4 /s at line beginning to be easily identified.
122 //// code
123 karl 1.1.2.35 There are still a number of KS_TODOs but they are mostly some questsions,
124 some internationalization (note that almost none of this environment is
125 completely internationlaized (see the decoders)) and a couple of questions
126 about the existing code or exactly what message should be sent. I argue
127 that the current TODOs should not stop checkin. Will remove after we
128 have tested couple of days and prove that there is no real need.
129 7. Added trace in Query response code if there is no objectPath since that
130 is really an error in the provider
131
|
132 karl 1.1.2.34
133 29 May 2014
134 1. Correct issue in cimcli that was causing rejects of open operations when
135 there were no filters requested. The standard default is WQL in cimcli
136 so we overrode this for the pull operations.
137 2. Clean up OpenQueryInstances in CIMMessage.h and CIMMessage.cpp.
138 3. Modify CIMOperationRequestDispatcher and
139 EnumerationContextTable.CreateContext to pass request rather than request
140 parameters.
141 4. Remove a number of diagnostic messages. Note that there are still some.
142 They are marked KS_TODO with with note to remove before checkin. They are
143 all traces and validations.
144 5. Correct error in inclusion of execinfo.h include in CIMOperationRequestDis.
145 found by Ji Lee.
146 6. Cleaned up execQuery and extended it for OpenQueryInstances. ExecQuery
147 tested but have not tested OpenQueryInstances yet. One more checkin and we
148 should be able to close 2,3 above (OpenQueryInstances).
|
149 karl 1.1.2.32
150 26 May 2014
151 1. Remove a number of diagnostics.
152 2. Make cache size in EnumerationContextTable dynamic and remove function
153 to set default values. MaximumEnums set from Dispatcher on constructor for
154 EnumerationContextTable and table size set to 1/2 of maximum number of
155 open contexts.
156 3. Clean up some statistics in EnumerationContextTable
|
157 karl 1.1.2.33 4. Correct error in query and queryLanguage parameters in cimcli in that
158 the default for queryLanguage is WQL but must be "" for pull operations.
159 If WQL found they set it to "".
|
160 karl 1.1.2.32
|
161 karl 1.1.2.31
162 9 May 2104
163 1. Converted several template functions in CIMOperationRequestDispatcher
164 to real functions
165 2. Added intermediate CIMMessage for CIMPullOperationRequestMessage since
166 all of the pulls have exactly the same data. This allowed us to
167 convert the template function processPullRequest from template to
168 real function
169 3. Added intermediate CIMMessage for CIMOpenOperationRequestMessage betweeh\n
170 all of the Opens and the CIMRequestData message. This will allow
171 common code to test all of these variables in all 6
172 CIMOperationRequestDispatcher handleOpen... functions
173 3. Added variable in CIMOperationRequestDispatcher handleEnqueue to make
174 releasing the original request conditional on return from the handle...
175 for open... and pull... handlers.
176 This allowed us to drop one place where we were creating a new request msg
177 for the delayed response and use the original request. This meant that
178 we could convert the issueOpenOrPullResponse from template function to
179 real function.
180 4. Moved open and pull constructors in CIMMessage.h to CIMMessage.cpp. NOTE:
181 openQueryInstances not moved. Will do when we get this funciton running.
182 karl 1.1.2.31 5. Added test for correct namespace in pull operations. Note that it is
183 never really used but since supplied with request, we test to be sure it is
184 same as open request namespace. added internationalized exception for this.
185 6. We should probably add same test to close just as an input parameter
186 validation.
187 6. Modified processPullRequest to close the enumeration context if
188 incorrect response type received. Before we just returned exception and
189 left context open for next request. Closing is much more logical since the
190 client really did send an erronous request.
191 7. Modified PullErrors to correctly test above changed in behavior.
192 8. Removed a significant number of diagnostic messages.
|
193 karl 1.1.2.33 9. Added query and querylanguage parameters to cimcli.
|
194 karl 1.1.2.31
|
195 karl 1.1.2.23
|
196 karl 1.1.2.30 4 May 2014
197 1. Clean interface with EnumerationContext timeout thread
198 2. Remove more diagnostics
199 3. Correct issue with enumerationContext Release that was causing
200 failures during nightly tests.
201
|
202 karl 1.1.2.29 2 May 2014
203 1. Remove Diagnostics from ResponseData.cpp
204 2. Correct setup and start of EnumerationContextTimerThread
205 3. Change limit on max number of enum contexts to be set from Dispatcher.
206 4. Increase size of hashtable for enum contexts.
207 5. Corrected issue causing failures in removeContext. Now runs through nightly
208 tests most of the time.
|
209 karl 1.1.2.28 1 May 2014
210 1. Cleanup of code and variable names.
211 2. Removed namespace variable from EnumerationContext. It was never used.
212 3. Added test for max number of open enumerations. Right now it is fixed
213 and set about 100. Note that this is tied to the size of hash table
214 so that there is a issue with more dynamic setting of max and the
215 hash table size.
216 4. Removed several unused functions.
217 5. Found one possible issue causing thread conflict but not the core one
218 yet.
219
220 30 April
221 1. Added exception response to dispatcher for Limits exceeded on
222 CreateEnumerationContext. Returns Exceeded server limits CIMException.
223 2. Fixed possible deadlock where we were not unlocking EnumerationContext
224 under error condition
225
226
|
227 karl 1.1.2.27 27 April - mergeout to head of tree
228 1. Mergeout of the code to the head of tree. Documented that mergeout in
229 the bug
230
231 27 April 2014
232 1. Clean up in a number of areas, mostly just code without changing
233 functionality.
234 2. Modified CIMMessage.h slightly to clarify names of the pull intermediate
235 messages (the common open and pull response);
236 3. Removed a number of diagnostics from CIMResponseData
237 4. Added new function to ResponseStressTestProvider to be able to terminate
238 output after a defined number of objects with a CIMException.
239 5. Removed a number of other TODOs
240 6. Corrected a couple of errors in the server message bundle.
241 7. Modified ResponseEncoder to only output the error for pull errors.
242 It was errounously outputting the EndOfSequence, EnumerationContext and
243 the error. This way, only the error is sent which is spec.
244 8. Removed number of TODOs and cleaned up a couple of others.
245 9. Added static tests for OpenEnumerateInstanceNames.
246 10. Modified a number of the temporary diagnostics to be permanent by
247 changing them to PEGASUS_DEBUG_ASSERTs
248 karl 1.1.2.27
|
249 karl 1.1.2.24 22 April 2014
250 1. Clean up some messages.
251 2. Still trying to find issue that randomly keeps messages in timeout
252 queue long after they have been closed and returned.
|
253 karl 1.1.2.25 3. Fixed memory loss error.
|
254 karl 1.1.2.26 4. Added zero reject for OperationTimeout because that effectively disables
255 the whole timeout mechanism so that enumerations could be started, never
256 completed by the client and would just hang around in memory. That
257 would be a significant DDOS attack (start very large enum with no
258 timeout) and never do pulls after open.
|
259 karl 1.1.2.25
|
260 karl 1.1.2.23 21 April 2014
|
261 karl 1.1.2.27 1. Reorganized code in CIMOperationRequestDecoder.cpp ...Iparam classes.
|
262 karl 1.1.2.23 Cleaned up the rejectCIMException code.
263 2. Cleanup some issues in the EnumerationContextTimeout Thread.
264 3. Fixed issue in pullop where it was using zero as default for default
265 operation timeout where it should have been NULL, the DMTF specified
266 default.
267 4. Enabled the test to reject 0 as operation timeout on a request and but
268 controled by a #define in CIMOperationRequestDispatcher.cpp
269 5. Added test for excessive new OperationContext entries in table.
270
271 18 April 2014
272 1. Internationalize remaining messages in CIMOperationRequestDecoder.cpp and
273 clean up the decoders for some of the messages so they use the
274 ...Iparam classes. Note that we did not clean up all of the old message
275 decoders, just the new ones and we did not clean up the Count message
276 since it is deprecated. TODO clean these up so the decoder is really
277 consistent.
278 2. Enabled the EnumerationContext Timer and removed code for alternate
279 non-thread solution.
|
280 karl 1.1.2.16
|
281 karl 1.1.2.21 17 April 2014
282 1. Reorganized the lock/unlock for handling the EnumerationContext because
283 there were thread windows (i.e. using the context after it was released
284 by other thread).
285 2. Modified code for the InteroperationTimerThread and deleted old code that
286 allowed this whole thread to by bypassed.
287 3. Internationalized a number of CIMOperationRequestDispatcher CIMException
288 response messages.
|
289 karl 1.1.2.22 4. Fixed stress tests in Client/tests/pullop that were getting timeout. The
290 issue was not the pull code but that fact that the idle client timeout
291 appears to be measureing to the end of the operation and the operation
292 was taking more than 40 seconds which is the default timeout. The idle
293 timeout takes no account of the fact that chunks are being regularly
294 delivered.
295 5. Several minor code changes.
296 NOTE: At this point the code passes extensive tests of the pull operations
297 both in OOP and none OOP mode (but with lots of diagnostics that may
298 be affecting timing).
|
299 karl 1.1.2.21
|
300 karl 1.1.2.20 10 April 2014
301 1. Fix issues causing failure with repeated pull operations.
302 2. Consolidate some of the CIMOperationRequestDispatcher.cpp template code
303
|
304 karl 1.1.2.19 7 April 2104
305 MergeOut of head of tree.
306
307 7 April 2014
308 1. Permanently fixed issue with getting correct number of objects on response.
309 This was introduced when did a temp fix to the OOP processing, in particular
310 removed the condition variable wait in the CIMOperationDispatcher handle
311 functions for opens and pulls. This moved the functionality to kick off
312 open and pull responses to the provider threads if there are not enough
313 objects when the handle is processed. Also added a variable to allow us to
314 test with either a) responses are required to satisfy the original
315 request size or) b, responses are required only to return some objects.
316 We will add statistics to see which of these works best.
317 2. Modified WsmProcessor EnumerationContext class since the name conflicts
318 with the pull Operation EnumerationContext class.
319 3. Created a new intermediate level of CIMMessage, the CIMPullResponseData
320 Message so that all of the open and Pull responses can use common code
321 since there are only two variables (endOfSequence and EnumerationContext)
322 and they are common across all the open and pull responses.
323 4. Removed a number of diagnostics.
324 5. Cleaned up the Dispcatcher so that the open and pulls have common response
325 karl 1.1.2.19 code and that code can be used from the dispatcher handle functions and
326 the provider response functions.
327
|
328 karl 1.1.2.18 31 March 2014 - Checkin
329 1. Fixed issues in OOP processing of pull operations, in particular
330 issues with cimxml output format when processed through the
331 *InternalXmlEncoder functions.
332 2. Clean up some of the internalXml functionality
333 3. Found issues causing timeout with a particular provider. The issue
334 is that the dispatcher and monitor end up using the same thread so the
335 condition variable in the dispatcher thread stops the monitor. Turned
336 off the conditionVariable in getCache for the moment which means that
337 we get number of responses for open... with 0 objects before the
338 providers can begin to respond. This is only for test.
339 4. Added some statistics for enumerations and display the statistics
340 when we close the server (same as cache statistics)
341
342 12 March 2014 - Mergeout and Mergein
343 1. Mergeout to head of tree for this date and mergein for patch update
344 to bug 9676
345 2. Extensions to pullop tests program and tests.
346 3. Added some diagnostics in looking for OOP issue.
347 4. Removed a number of diagnostics messages and cleaned up code in
348 dispatcher to simplify pull operation processing.
349 karl 1.1.2.18
|
350 karl 1.1.2.17 15 December 2013
351 1. Mergeout and mergein up to 15 December 2013
352 2. Clean up issues from tests documented in bug 9676 last week.
353 3. Clean up some code in dispatcher
354 4. Remove the filter function from ResponseStressc++Provider.
355
|
356 karl 1.1.2.16 21 November 2013
357 1. Mergeout from head of tree to 21 November 2013.
|
358 karl 1.1.2.15
359 18 November 2013
360 1. Cleanup of a bunch of minor errors and completion of all of the code for
361 the openQueryInstances except for the PullInstances in Dispatcher and
362 the aggregator function.
363 2. OpenqueryInstances added to cimcli.
|
364 karl 1.1.2.14
365 13 October 2013 CVS branch update.
366 1. Integrated bug 9786 into the branch. Note that we need to test the
367 generated statistics.
368 2. Mergeout executed to update to head of tree as of 8:00 am 13 October 2013.
369 3. Cleaned up several errors in OOP processing. Note that there is at least
370 one issue left when we to a pull on ManagedElement in at least one of the
371 namespaces.
372 4. Cleaned up some of the outstanding diagnostic code
373 5. Generally passes all tests except for one test of pullop where it is trying
374 to pull enum instances CIM_ManagedElement from a particular namespace.
|
375 karl 1.1.2.12
|
376 karl 1.1.2.16 NOTE: I did not make comments here for changes in October despite the fact
377 that I did 2 mergouts, number of fixes, and a mergein.
378
|
379 karl 1.1.2.12 30 September 2013 - CVS Update
380 Mergeout head of tree up to 29 September 2013.
|
381 karl 1.1.2.10
382 29 September 2013. CVS update.
383 1. Modified calls to statisticalData.cpp to a) directly call with request
384 type, b) incorporate the open, pull, etc. messages. However, since these
385 are not part of the CIM class, we must do something special with them.
|
386 karl 1.1.2.13 See bug 9785 for full solution to this issue.
|
387 karl 1.1.2.10 2. Corrected OOP interface to enable new flag to indicate internal operations
388 and set host, etc.
389 3. Add code to CQLOperationsDispatcher and WQLOperationDispatcher to clean
390 up CIMResponseDataCounter after filtering.
391 4. Modified ProviderAgent to set Host info for some pull operations.
392 5. Added new flag to CIMBinMsgSerializer and Deserializer.
|
393 karl 1.1.2.8
|
394 karl 1.1.2.9 17 September 2013 CVS update (Actually two different updates over 3 days)
395 1. Clean up some issues in CIMMessage.h and CIMMessage.cpp
396 2. Extend OpenExecQuery to WQL and CQL processors but return not complete
397 3. Remove memory leak in EnumerationContext and EnumerationContextTable
398 handling.
399 4. Created template functions for much of the pull operations.
400 5. Reversed order of queryLanguage and query (and changed names to match
401 execQuery) in client and server. Note that these are the execQuery
402 WQL and CQL filters and NOT FQL filters.
403 6. Some code cleanup in dispatcher
404 7. Today, passes all tests in pullop but issue in alltests. For some reason
405 not finding CIMObjectManager instance. Also, leaves enumeration contexts
406 if client terminates since cleanup thread not operating.
|
407 karl 1.1.2.11 8. XML from OOP not correctly processed.
|
408 karl 1.1.2.9
|
409 karl 1.1.2.8 14 September 2013 CVS update
410 Merged out up to 25 August. Cleaned up all operations and standardized code.
411 At this point the non pull operations code is in a set of templates but the
412 pull is not yet.
413 Fixed a significant number of problems so that it appears that the operations
414 except for OpenExecQuery run stably, at least with the pullop test program.
415 Note that there is a problem in that the Interop control provider is not
416 returning its singleton wbemserver object for some reason. Causes a test
417 failure
|
418 karl 1.1.2.7
419 Fixed for 16 June CVS Update
420 1. Cleaned up the enumerationContext and Table release functions and tested
421 to confirm that we do not lose memory in either normal sequences or
422 sequences that close early. Cleaned up pullop and added more tests
|
423 karl 1.1.2.8 Taged Before: PREAUG25UPDATE and after POSTAUG25UPDATE
|
424 karl 1.1.2.4
425 Fixed for 9 June CVS update
426 1. Cleaned up code for OpenQueryInstances. Note that this is incomplete.
427 No support in WQL or CQL Operations
428 2.
429
430 What was fixed for 5 June checkin.
|
431 karl 1.1.2.3 1. Extended ResponseTest MOF for for both CMPI and C++ subclasses
432 2. Fixed issues with pullop.
433 3. Fixed temp issue with CIMResponseData size by putting in mutex. That
434 is not a permanent fix but it gets around issue probably in the control
435 of the move logic that meant counts were off.
436 4. Fixed issues in Dispatcher so that associator code works. Still messy
437 code in the dispatcher.
438 5. Changed name of Enumerationtable.h & cpp to EnumerationContextTable.*
439 6 Changed name of ResponseStressTest module, classes, etc.
440
441 TAG: TASK_PEP317_5JUNE_2013_2
442
|
443 karl 1.1.2.2 2 June 2013
|
444 karl 1.1.2.1
445 Issues - KS
446
|
447 karl 1.1.2.15 - Still way to many TODO and KS comments and KS_TEMPS. Removing bit by bit.
|
448 karl 1.1.2.1
|
449 karl 1.1.2.15 - Runtime variable connection for the config parameters not installed. That
450 has been made into a separate bug (see bug 9819)
|
451 karl 1.1.2.1
452 5. Issue with the threaded timer. For some reason during tests it
453 eventually calls the timer thread with trash for the parm (which is
454 pointer to the EnumerationTable object). Caught because we do a valid
455 test at beginning of the function.
456
|
457 karl 1.1.2.2 6. Still using the templates in CIMOperationRequestDispatcher to simplify
458 the handle... processing.
459
460 7. I think I have a way around the double move of objects in the
461 EnumerationContext so that the outputter will just take a defined number
462 of objects directly from the gathering cache and save the second move.
463
|
464 karl 1.1.2.15 8. Not yet passing all tests but getting closer now. The major test that is
465 causing an error today is the execution of a full enumeration with the
466 forceProviders = true. This causes a client timeout sometimes.
|
467 karl 1.1.2.2
468
469
|
470 karl 1.1.2.1 ===========================================
471
472 OVERVIEW:
473
474 The operation extensions for pull operations defined in the DMTF specification
475 DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
476 including Client and Server.
477
478 These operations extend the CIM/XML individual operations to operation
479 sequences where the server must maintain state between operations in a
480 sequence and the client must execute multiple operations to get the full
481 set of instances or instance paths.
482
483 The following new CIM/XML operations as defined in DSP0200 are included;
484
485 -OpenEnumerateInstances
486 -openEnumerateInstancePaths
487 -OpenReferenceInstances
488 -OpenReferenceInstancePaths
489 -OpenAssociatiorInstances
490 -OpenAssociatorInstancePaths
|
491 karl 1.1.2.16 -OpenQueryInstances
|
492 karl 1.1.2.1 -PullInstancesWithPath
493 -PullInstancePaths
|
494 karl 1.1.2.16 -PullInstances
|
495 karl 1.1.2.1 -CloseEnumeration
496 -EnumerationCount
|
497 karl 1.1.2.2 OpenExecQuery
|
498 karl 1.1.2.1
499 The following operations have not been implemented in this version of Pegasus:
500
501 -OpenQueryInstances
502
503 The following limitations on the implementation exist;
504
505 1. The filterQueryLanguage and filterQuery parameters are processed by
506 the Pegasus client but the server returns error if there is any data in
|
507 karl 1.1.2.2 either parameter. This work does not include the development of the
508 query language. Note that a separate effort to extend Pegasus to use
509 the DMTF FQL query language is in process.
|
510 karl 1.1.2.1
511 2. The input parameter continueOnError is processed correctly by the client
512 but the Pegasus server only provides for false since the server does not
513 include logic to continue processing responses after an error is
514 encountered.
515 This is consistent with the statement in the specification that use of
516 this functionality is optional and the fact that the DMTF agrees that all
517 of the issues of continuing after errors have not been clarified.
518
519 3. The operation enumerationCount is not processed by the server today since
520 a) really getting the count would be the same cost as the corresponding
521 enumeration, b) the server does not include a history or estimating
522 mechanism for this to date.
|
523 karl 1.1.2.2 NOTE: After a through review as part of the development of the next version
524 of CMPI we have concluded that this operation is probably not worth the
525 effort. Since it is optional, Pegasus will only return the unknown status
526 at this point
|
527 karl 1.1.2.1
528 Since the concept of sequences of operations linked together (open, pull, close)
529 is a major extension to the original CIM/XML operation concept of completely
530 independent operations several new pieces of functionality are implemented
531 to control interOperationTimeouts, counts of objects to be returned, etc.
532
533 TBD - Review this
534
535 CLIENT
536
537 The new operations follow the same pattern as the APIs for existing operations
538 in that:
539
540 1. All errors are handled as CIMException and Exception
541
542 2. The means of inputting parameters are the same except that there are
543 significantly more input parameters with the open operations and for the
544 first time operations return parameters as well as objects in the
545 response. Specifically the open and pull operations return values for
546 enumerationContext which is the identity for a pull sequence and
547 endOfSequence which is the marker the server sends in open and pull
548 karl 1.1.2.1 responses when it has no more objects to send.
549
550 The significant differences include:
551
552 1. Processing of parameters on responses (i.e. the endOfSequence and
553 enumerationContext parameters are returned for open and pull operations).
554
555 2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
556 cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
557 client api.
558
559 3. The association and reference operations ONLY process instances. They do
560 not include the capability to return classes like reference and associator
561 do and therefore return CIMInstance rather than CIMObject.
562
563 4. Paths are returned in all cases (i.e OpenEnumerateInstances and
564 PullInstancesWithPath where they were not with EnumeratInstances.
565
566 5. The client must maintain state between operations in a sequence (using
567 the enumerationContext parameter).
568
569 karl 1.1.2.1 TBD- Are there more differences.
570
571
572 SERVER
573
574 The Pegasus server attempts to always deliver the requested number of objects
575 for any open or pull request (the specification allows for the server to
576 deliver less than the requested number of objects and specifically to return
577 zero objects on open). We felt that it was worth any extra cost in processing
578 to provide the client with exactly what it had requested.
579
580 The pegasus server always closes an enumeration sequence upon receipt of any
581 error from the providers, repository, etc. Therefore the server will reject
|
582 karl 1.1.2.2 any request that has continueOnError = true;
583
584 Expansion to allow the continue on error may be added in a future version.
585 In any case, the whole purpose of the continue on error is really to allow
586 input from good providers to be mixed with providers that return errors so
587 that generally this would mean simply changing the logic in the return mechanism
588 to not shutdown when an error is received from any given provider.
589
590 Generally we do not believe that the providers need to do much more in the
591 future to support the continueOnError other than possibly allowing the provider
592 to continue processing after it has received an error.
|
593 karl 1.1.2.1
594 PROVIDERS
595
596 This implementation requires NO changes to the existing providers. The
597 provider APIs operate just as they do with the original operations.
598
599 Because the server processing is different however, there may be some
600 behavior differences primarily because the client now controls the speed of
601 delivery of objects.
602
603 In previous versions of Pegasus, the server attempts to deliver objects as
604 rapidly as then can be put on the network. In the case of HTTP chunked requests
605 they are delivered in chunks of about 100 objects. The primary delay for the
606 providers was the processing of each segment through the server. The server
607 is blocked so that no other segment can proceed through the server until that
608 segment is processed and sent on the network.
609 In the case of non-chunkedresponses, they are completely gathered in the serve
610 and then delivered as one non-chunked response. There were no delays for the
611 providers, just lots of possible memory use in the server.
612
613 The responses from providers (delivered through the deliver(...) interface are
614 karl 1.1.2.1 gathered into segments of about 100 objects and this group of objects is moved
615 through the server to be delivered to the client.
616
617 However with the inclusion of the pull operations, The segments of objects
618 from the providers are cached in the server response path until the
619 maxObjectCount for that request (open or pull) and that number returned in a
620 non-chunked response. Thus, if the client is slow to issue pull requests,
621 the providers might be delayed at some point to reduce memory usage in the
622 server (the delay appears as slow response tothe deliver operation).
623
624 In other words, the time to process large sets of responses from the provider
625 now depends on the speed of handling the client.
626
627 It is important to remember in developing providers that the Pegasus server
628 can most efficiently process responses if they are passed from the provider
629 to the server individually or in small arrays of objects rather than the
630 provider gathering very large arrays of objects and sending them to the
631 server.
632
|
633 karl 1.1.2.2 NEXT GENERATION PROVIDERS
634 KS_TODO
635
|
636 karl 1.1.2.1 CONFIGURATION PARAMETERS
637
638 The server includes several configuration parameters to set limits on the
639 processing of pull operations. All of these configuration parameters are
640 compile time parameters rather than runtime.
641
642 1. Maximum value of minimum interoperation time. This parameter defines the
643 maximum time allowed between the return of an open or pull response and
644 the receipt of the next pull or a close operation before the server may
645 close the enumeration. The specification allows the server to set a
646 maximum interoperation time and refuse open requests that with requested
647 operationTimeout greater than that time.
648 CIM_ERR_INVALID_OPERATION_TIMEOUT
649
650 This value is set with the Pegasus environment variable
651 PEGASUS_PULL....
652
653 2. Maximum objects returned in a single open or pull operation. The server
654 can set a maximum limit on the number of objects that can be returned in
655 a single open or pull oepration with the maxObjectCount parameter.
656
657 karl 1.1.2.1 3. Whether the server allows 0 as an interoperation timeout value. The value
658 zero is s special value for the interoperationTimeout in that it tells the
659 server to not timeout any enumeration sequence.
660
661 With this value for interoperationTimeout, the only way to close an
662 enumeration sequence is to complete all of the pulls or issue the close.
663 If for some reason the sequence is not completed, that enumeration context
664 would remain open indefinitly. Since in Pegasus any open enumeration
665 context uses resources (the context object and any provider resposnes that
666 have not yet been issued in a response) it would appear that most
667 platforms would not want to allow the existence of enumeration contexts
668 that cannot be closed by the server.
669
670 4, maximum consecutive pull requests with 0 maxObjectCount. The use of the
671 pull operation with maxObjectCount set to zero could be used to keep an
672 enumeration context open indefinitly (this tells the server to restart the
673 interoperationTimeout but not send any objects in the response). Therefore the
674 specification allows for the server setting maximum limits on this behavior
675 and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
676 exceeded.
677 Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
678 karl 1.1.2.1 a non-zero count resets this counter.
679
680 KS-TBD - Is this really logical since we can still block by just issuing
681 lots of zero request and an occansional request for one object.
682
683 Pegaus sets the value of this limit to 1000 and allows the implementer to
684 modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
685
686 5. Default operationTimeout -
687
688 The default of this parameter is to refuse operat
689
690 In the current release of Pegasus these are all compile time parameters.
|
691 karl 1.1.2.11
692
693 NOTES On working with task branch.
694
695 Merge out Process
696
697 To keep our TASK branch in sync with the current head of tree we need
698 to do a regular merge out. the TaskMakefile contains the makefile
699 procedures to do this efficiently. NOTE: Following these procedures is
700 important in that you are merging out new material each time you do
701 the merge out. If you were just to repeatedly merge out, you would be
702 merging previously merged changes a second time causing a real mess.
703
704 Start with new directory and put TaskMakefile above pegasus (needed so you
705 have this file for the initial operations.
706
707 make -f TaskMakefile branch_merge_out BNAME=PEP317-pullop ## takes a long time
708
709 This checks out current head, merges it into task branch and sets tags
710 for the mergeout. Note that at the end of this step this work is
711 part of the TASK... branch.
712 karl 1.1.2.11
713 NOW check for conflicts, errors, etc. that resulted from the merge.
714 Look for conflict flags, compare the results (I use linux merge as a
715 good graphic compare tool) and build and test. When you are satisfied
716 that the merge out is clean, you can commit the results to the TASK...
717 branch
718
719 To commit the work to this into Task branch
720
721 make -f mak/TaskMakefile branch_merge_out_commit BNAME=PEP317-pullop
722
723 or manually commit and finish as follows
724
725 cvs commit
726 make -f mak/TaskMakefile branch_merge_out_finish BNAME=PEP317-pullop
727
728 ## This last step is important since it cleans up temporary tags to prepare
729 you for the next checkout
730
731 COMPARE TASKBRANCH WITH HEAD
732
733 karl 1.1.2.11 In a new pegasus work space do same as above for merge out.
734
735 make -f TaskMakefile BNAME=PEP317-pullop
736
737 This produces a result which is all of the head merged into the branch.
738 A diff of this is all the new changes to the head of tree that you will
739 include into the merge.
740
|