1 karl 1.2 Using the CIM/XML Pull Operations
2
3 STATUS
4
5 <<< The TODO section is being maintained during the review and checkin process
6 to keep track of problems, errors, notes, etc. Must be deleted before
7 release or reduced to those things that are considered FUTURE and not for
8 this release. Please feel free to add notes, etc in this
9 section as you review/test.>>>>>>
10
11 TODO list:
12 1. Binary operation from OOP. Need to add counter to binary
13 protocol to be able to count objects in response. Generates
14 warnings in things like messageserializer and does not work with
15 OOP right now. Fixed by converting to XML. Concluded that we do not
16 need to do this. The binary response is not really used often
17 in the current environment So double mapping it is not a major issue.
18 Leave this as FUTURE
19 2. DONE OpenExecQuery - Code is incomplete in that it does not include the
20 return from the exec query function to the aggregator yet. Code for Pull
21 part of OpenQueryInstancesRequest a) should be part of the common
22 karl 1.2 CIMOperationRequestDispatcher execCommon code. DONE, not tested. 29 May
23 30 May tested with simple cli calls.
24 3. DONE The changes to WQLCIMOperationRequestDispatcher and CQL... for handling
25 pull not completed so we feed the responses back to the EnmerationContext
26 queues. DONE Not tested 29 May 30 May DONE, some testing done
27 4. Minor TODOs, diagnostics, etc. still in the code. Reduced almost to
28 none now. We are leaving some in as PEGASUS_DEBUG
29 5. Extension to avoid double move of objects in CIMResponseData (one
30 into enumerationContext queue and second to new cimResponseData for
31 response. Want to avoid second move by extending Open/Pull response
32 messages to include count and CIMResponse data to count objects out
33 of queue when converting (avoids the second move). This would mean
34 extending the output writers to allow the count field to be supplied
35 so they would only create output for up to the count supplied.(Make
36 this future beyond bug 9676). This is fairly extensive because it
37 extends beyond CIMResponseData to SCMO and XML writers where the
38 XmlWriters used by encodeXmlResponse would have to have counters
39 added. Then instead of copying on getCache we would simply pass the
40 cache and count on and the writer would take and remove.
41 6. Add more static tests (currently only OpenEnumerateInstances and
42 OpenEnumerateInstanceNames covered).
43 karl 1.2 Low priority because most of this except exact XML covered in other
44 tests. note that we can only add very limited tests since a) no
45 way to test multiple operations in sequence(i.e. open/pull) and
46 cannot be sure how many responses will come back in all cases even
47 for open. The only reason for most of these tests is to assure that
48 we do not change XML inadvertendtly in future with change that is
49 compatible to both our server and client but breaks other clients or
50 servers. In any case we can do only minimal tests because most real
51 testing involves both open and pull and the static tests are by their
52 nature single operation.
53 8. Correct issue between operations and HTTP where we are sending
54 trailers with exceptions. Modify response mechanisms so that we
55 set non-chunked for all responses where we send error responses to
56 avoid the trailers. NOTE: There should be now a bug on this in general
57 where we would want to send an initial error without the trailer. Should
58 have always done that.
59 9. It would be more efficient in dispatcher to always use exception for
60 rejects and change the _reject functions so that they never return
61 when they reject. This is VERY LOW PRIORITY and primarily saves
62 a few lines of code in the reject functions and their calls. Means we
63 would code.
64 karl 1.2 _rejectIfEnumerationToBroad(...);
65 _rejectIfThisParameterIncorrect(...);
66
67 instead of
68 if (_rejectIfEnum...)
69 {
70 return true
71 }
72 It would mean that the method trace for the handlers would not return
73 an exit if we rejected. VERY LOW PRIORITY. Possibly FUTURE. No behavior
74 change, just more compact source code but it messes with the method
75 trace logic.
76 10. There are still a couple of template functions around the task of
77 distributing requests to the multiple providers.
78 In fact there are two similiar but different templates for
79 the associators(i.e. assoc, ref, etc.) functions and the parallel
80 openAssoc functions. It would be nice to consolidate that logic and
81 further to try to create a non-template form for those functions. LOW
82 PRIORITY and not for bug 9676 ACTION: Define in new bug
83 11. External runtime variables. bug 9676 leaves these variables as
84 internal to CIMOperationRequestDispatcher.cpp.
85 karl 1.2 See bug 9819 for the changes to externalize these variables, both
86 the runtime and buildtime variables. The Work group discussion
87 documents what the variables are and how they tie to statistics info.
88 See the OpenPegasus Wiki, Pull operations work group.
89 12. Trace functions in CIMResponseData,h & cpp should be PEGASUS_DEBUG.
90 Same for trace function in EnumerationContext and
91 EnumerationContextTable
92 13.Question. We added trace in CIMRequestOperationDispatcher.cpp if
93 query enum is recevied with no object path. However, since this is
94 a provider problem should we be doing something other than a trace
95 (ex. log) since traces are often not really used and they do generate
96 lots of data whereby something like this could get lost. Also, need
97 to review what level of trace if we keep trace.
98 14. Note that the build-a-class argument on OpenQueryInstances always
99 returns error since we do not build the class. We propose to deprecate
100 this in DMTF and not implement in in OpenPegasus. TODO add a bug
101 defining this limitation. This is not a TODO, just a warning for
102 the future. TODO add bug about this and add to release notes
103 15. Clarify trace of completion of each enumeration sequence.
104 16. Statistics keeping for open, etc. through cimperf.
105
106 karl 1.2 NEXT TASKS:
107 a. finish 8, 12 above. Test against cimperf since there was a bug
108 in cimperf from before.
109
110 26 August 2014
111 1. Add flag to pullop to hide warnings when response smaller than requested
112 size. This is because the default operation is to return what exists rather
113 than waiting for a full response to match the MAXOBJECTCNT. Note that
114 there is a flag in CIMOperationRequestDispatcher that modifies this
115 behavior.
116 2. Modify displays in pullop to display less info during poststarttests.
117
118 20 August 2014
119 1. Update to greatly reduce the number of KS_TODO comments. We have reduce
120 the number of these comments to about 50 total of which most are comments
121 to remove diagnostics.
122 2. Internationlize a number of messages in XmlReader, etc. This completes
123 internationalization of messages added for pull operations.
124 3. Added some detailed messages to cimcli to show what is happening in
125 the multiple operations.
126 4. Some efficiency changes (ex. remove String::EMPTY usage)
127 karl 1.2
128 18 August 2014
129 1. Update to remove some TODOs.
130 2. Mergeout to the task branch. Note that for some reason there had bee
131 a descrepency in the Pegasus/Server/Tests/TLS last bug fix not getting
132 into the previous mergeout
133
134 14 August 2014
135 1. Issue with the pullop test in some systems causing an HTTP 413 error (
136 too much data). We confirmed that this is probably actually a limit in the
137 client with enumerate instances (the response is simply to big in the case
138 where we are asking it to deliver 100,000 responses instances on some systems)
139 Modified the test for the stress test to be 30,000 instances and added a
140 manual test testSizeLimit to the make file to try to deliver up to 600,000
141 instances to determine limits to the environment.
142
143 Note that this is not actually a failure of the pull operations. In fact
144 where cimcli ei ... fails with 600,000 instances in the response cimcli pei
145 works correctly.
146
147 Update the Client/tests/pullop/Makefile stresstests to hopefully pass the
148 karl 1.2 nightly test on more systems.
149
150 2. Eliminated a number of KS_TODOs and code that we had documented out with
151 //// . Note that there are still a number of the KS_TODOs but most of them
152 are really isolated around
153
154 19 August 2014
155 1. Mergeout to head of tree. We picked up one issue that for some reason
156 the cvs mergeout was missing, a change to Server/tests/TLS...
157
158 19 August 2014
159 1. Remove a number of KS_TODO notes either by cleaning them up or by
160 change the keyword to FUTURE if they are things that we might consider
161 for the future
162 2. Some efficiency improvements in dealing with strings primarily get
163 rid of use of String::EMPTY in the new code.
164
165 3. Removed some diagnostic trace output that was used during development.
166
167 13 August 2014
168 1. Mergeout to head of tree.
169 karl 1.2
170 11. August 2014
171 1. Correct error in compile in windows (bad reference to sleep(...)
172
173 29 July 2014
174 1. Cleanup functionality in the unittests for EnumerationContext and
175 its table.
176
177 23 July 2014
178 1. Mergeout from head of tree to Task branch
179
180 21 July 2014
181 1. Removed strings.h from pullop.cpp
182 2. Minor cleanup per testing
183
184 17 July 2014
185 1. Correct error in CIMOperationRequestDispatcher.cpp OpenQueryInstances
186 when PEGASUS_DISABLE_QUERY set.
187 2. Correct issue in diagnostic code in CIMResponseData.cpp that used
188 snprintf function not available in windows.
189 3. Clean up some miscellaneous code in pullop and Makefile.
190 karl 1.2 4. Minor change to PullErrosrs.
191
192 2 July 2014
193 1. Corrected error in the new getoopt functions that caused it to fail
194 in build and also in it unit test with the --help option.
195
196 28 June 2014
197 1. Corrected issue where ZOs was getting errors from pullop because we
198 used getopt to parse command line and the linux getopt is not behavior
199 consistent with ZOs. Note that the ZOs getopt matches the posix expect and
200 the linux one has extensions (notably it permutes non option arguments to
201 the end of the arguments array. The easiest and most logical was to move
202 the whole parser to use the internal getoopt which is what we should be
203 using all the time in any case.
204 18 June 2014
205 1. Correct errors in pullop and CMIMessage.cpp identified by Ji Lee as comments 58,59
206 2. Correct error in Makefile identified by Ji Lee as bug comment 60
207
208 14 June 2014
209 1. Mergeout from head of tree.
210
211 karl 1.2 12 June 2014
212 1. Implement the basic code for the CIMOperationDispatcher portion of bug 9819
213 Note that the CIMConfig calls are disabled with comments.
214 2. Corrected a number of differences in const declaration and definition for
215 continueOnError and maxObjectCount in CIMClient. This corrects issue
216 documented in bug 9676 comment 53
217 3. Corrected some documentation for Pull operations in CIMClient.
218 4. Corrected similar issue In EnumerationContext class. See comment 54
219 5. Removed ExecInfo.h header from Client/tests/PullErrors/PullErrors.cpp
220
221 30 May 2014
222 1. Correct issue in UintArgs. Modified to use smart pointers, consistent
223 with public interfaces. Tested
224 2. Complete code for OpenQueryInstances and PullInstances so that the complete
225 sequence works for both the server and client (i.e. 2, 3 above done
226 except for more tests in pullop).
227 3. Fix issue in cimcli where the OpenQueryInstances was outputting a path
228 component where there is no path transmitted from the server.
229 4. Moved large block of common code in CQLOperationRequestDispatcher.cpp
230 and WQLOperationRequestDispatcher.cpp to common function
231 handleQueryRequestCommon in CIMOperationRequestDispatcher.cpp. These
232 karl 1.2 handlers now reduced to just the parser processing and minimal error
233 checks.
234 5. Cleaned up some code in EnumerationContextTable.CreateContext so only
235 single interface for all requests.
236 6, Commented out a number of PEG_TRACE messages. Note that they are
237 commented with 4 /s at line beginning to be easily identified.
238 //// code
239 There are still a number of KS_TODOs but they are mostly some questsions,
240 some internationalization (note that almost none of this environment is
241 completely internationlaized (see the decoders)) and a couple of questions
242 about the existing code or exactly what message should be sent. I argue
243 that the current TODOs should not stop checkin. Will remove after we
244 have tested couple of days and prove that there is no real need.
245 7. Added trace in Query response code if there is no objectPath since that
246 is really an error in the provider
247
248
249 29 May 2014
250 1. Correct issue in cimcli that was causing rejects of open operations when
251 there were no filters requested. The standard default is WQL in cimcli
252 so we overrode this for the pull operations.
253 karl 1.2 2. Clean up OpenQueryInstances in CIMMessage.h and CIMMessage.cpp.
254 3. Modify CIMOperationRequestDispatcher and
255 EnumerationContextTable.CreateContext to pass request rather than request
256 parameters.
257 4. Remove a number of diagnostic messages. Note that there are still some.
258 They are marked KS_TODO with with note to remove before checkin. They are
259 all traces and validations.
260 5. Correct error in inclusion of execinfo.h include in CIMOperationRequestDis.
261 found by Ji Lee.
262 6. Cleaned up execQuery and extended it for OpenQueryInstances. ExecQuery
263 tested but have not tested OpenQueryInstances yet. One more checkin and we
264 should be able to close 2,3 above (OpenQueryInstances).
265
266 26 May 2014
267 1. Remove a number of diagnostics.
268 2. Make cache size in EnumerationContextTable dynamic and remove function
269 to set default values. MaximumEnums set from Dispatcher on constructor for
270 EnumerationContextTable and table size set to 1/2 of maximum number of
271 open contexts.
272 3. Clean up some statistics in EnumerationContextTable
273 4. Correct error in query and queryLanguage parameters in cimcli in that
274 karl 1.2 the default for queryLanguage is WQL but must be "" for pull operations.
275 If WQL found they set it to "".
276
277
278 9 May 2104
279 1. Converted several template functions in CIMOperationRequestDispatcher
280 to real functions
281 2. Added intermediate CIMMessage for CIMPullOperationRequestMessage since
282 all of the pulls have exactly the same data. This allowed us to
283 convert the template function processPullRequest from template to
284 real function
285 3. Added intermediate CIMMessage for CIMOpenOperationRequestMessage betweeh\n
286 all of the Opens and the CIMRequestData message. This will allow
287 common code to test all of these variables in all 6
288 CIMOperationRequestDispatcher handleOpen... functions
289 3. Added variable in CIMOperationRequestDispatcher handleEnqueue to make
290 releasing the original request conditional on return from the handle...
291 for open... and pull... handlers.
292 This allowed us to drop one place where we were creating a new request msg
293 for the delayed response and use the original request. This meant that
294 we could convert the issueOpenOrPullResponse from template function to
295 karl 1.2 real function.
296 4. Moved open and pull constructors in CIMMessage.h to CIMMessage.cpp. NOTE:
297 openQueryInstances not moved. Will do when we get this funciton running.
298 5. Added test for correct namespace in pull operations. Note that it is
299 never really used but since supplied with request, we test to be sure it is
300 same as open request namespace. added internationalized exception for this.
301 6. We should probably add same test to close just as an input parameter
302 validation.
303 6. Modified processPullRequest to close the enumeration context if
304 incorrect response type received. Before we just returned exception and
305 left context open for next request. Closing is much more logical since the
306 client really did send an erronous request.
307 7. Modified PullErrors to correctly test above changed in behavior.
308 8. Removed a significant number of diagnostic messages.
309 9. Added query and querylanguage parameters to cimcli.
310
311
312 4 May 2014
313 1. Clean interface with EnumerationContext timeout thread
314 2. Remove more diagnostics
315 3. Correct issue with enumerationContext Release that was causing
316 karl 1.2 failures during nightly tests.
317
318 2 May 2014
319 1. Remove Diagnostics from ResponseData.cpp
320 2. Correct setup and start of EnumerationContextTimerThread
321 3. Change limit on max number of enum contexts to be set from Dispatcher.
322 4. Increase size of hashtable for enum contexts.
323 5. Corrected issue causing failures in removeContext. Now runs through nightly
324 tests most of the time.
325 1 May 2014
326 1. Cleanup of code and variable names.
327 2. Removed namespace variable from EnumerationContext. It was never used.
328 3. Added test for max number of open enumerations. Right now it is fixed
329 and set about 100. Note that this is tied to the size of hash table
330 so that there is a issue with more dynamic setting of max and the
331 hash table size.
332 4. Removed several unused functions.
333 5. Found one possible issue causing thread conflict but not the core one
334 yet.
335
336 30 April
337 karl 1.2 1. Added exception response to dispatcher for Limits exceeded on
338 CreateEnumerationContext. Returns Exceeded server limits CIMException.
339 2. Fixed possible deadlock where we were not unlocking EnumerationContext
340 under error condition
341
342
343 27 April - mergeout to head of tree
344 1. Mergeout of the code to the head of tree. Documented that mergeout in
345 the bug
346
347 27 April 2014
348 1. Clean up in a number of areas, mostly just code without changing
349 functionality.
350 2. Modified CIMMessage.h slightly to clarify names of the pull intermediate
351 messages (the common open and pull response);
352 3. Removed a number of diagnostics from CIMResponseData
353 4. Added new function to ResponseStressTestProvider to be able to terminate
354 output after a defined number of objects with a CIMException.
355 5. Removed a number of other TODOs
356 6. Corrected a couple of errors in the server message bundle.
357 7. Modified ResponseEncoder to only output the error for pull errors.
358 karl 1.2 It was errounously outputting the EndOfSequence, EnumerationContext and
359 the error. This way, only the error is sent which is spec.
360 8. Removed number of TODOs and cleaned up a couple of others.
361 9. Added static tests for OpenEnumerateInstanceNames.
362 10. Modified a number of the temporary diagnostics to be permanent by
363 changing them to PEGASUS_DEBUG_ASSERTs
364
365 22 April 2014
366 1. Clean up some messages.
367 2. Still trying to find issue that randomly keeps messages in timeout
368 queue long after they have been closed and returned.
369 3. Fixed memory loss error.
370 4. Added zero reject for OperationTimeout because that effectively disables
371 the whole timeout mechanism so that enumerations could be started, never
372 completed by the client and would just hang around in memory. That
373 would be a significant DDOS attack (start very large enum with no
374 timeout) and never do pulls after open.
375
376 21 April 2014
377 1. Reorganized code in CIMOperationRequestDecoder.cpp ...Iparam classes.
378 Cleaned up the rejectCIMException code.
379 karl 1.2 2. Cleanup some issues in the EnumerationContextTimeout Thread.
380 3. Fixed issue in pullop where it was using zero as default for default
381 operation timeout where it should have been NULL, the DMTF specified
382 default.
383 4. Enabled the test to reject 0 as operation timeout on a request and but
384 controled by a #define in CIMOperationRequestDispatcher.cpp
385 5. Added test for excessive new OperationContext entries in table.
386
387 18 April 2014
388 1. Internationalize remaining messages in CIMOperationRequestDecoder.cpp and
389 clean up the decoders for some of the messages so they use the
390 ...Iparam classes. Note that we did not clean up all of the old message
391 decoders, just the new ones and we did not clean up the Count message
392 since it is deprecated. TODO clean these up so the decoder is really
393 consistent.
394 2. Enabled the EnumerationContext Timer and removed code for alternate
395 non-thread solution.
396
397 17 April 2014
398 1. Reorganized the lock/unlock for handling the EnumerationContext because
399 there were thread windows (i.e. using the context after it was released
400 karl 1.2 by other thread).
401 2. Modified code for the InteroperationTimerThread and deleted old code that
402 allowed this whole thread to by bypassed.
403 3. Internationalized a number of CIMOperationRequestDispatcher CIMException
404 response messages.
405 4. Fixed stress tests in Client/tests/pullop that were getting timeout. The
406 issue was not the pull code but that fact that the idle client timeout
407 appears to be measureing to the end of the operation and the operation
408 was taking more than 40 seconds which is the default timeout. The idle
409 timeout takes no account of the fact that chunks are being regularly
410 delivered.
411 5. Several minor code changes.
412 NOTE: At this point the code passes extensive tests of the pull operations
413 both in OOP and none OOP mode (but with lots of diagnostics that may
414 be affecting timing).
415
416 10 April 2014
417 1. Fix issues causing failure with repeated pull operations.
418 2. Consolidate some of the CIMOperationRequestDispatcher.cpp template code
419
420 7 April 2104
421 karl 1.2 MergeOut of head of tree.
422
423 7 April 2014
424 1. Permanently fixed issue with getting correct number of objects on response.
425 This was introduced when did a temp fix to the OOP processing, in particular
426 removed the condition variable wait in the CIMOperationDispatcher handle
427 functions for opens and pulls. This moved the functionality to kick off
428 open and pull responses to the provider threads if there are not enough
429 objects when the handle is processed. Also added a variable to allow us to
430 test with either a) responses are required to satisfy the original
431 request size or) b, responses are required only to return some objects.
432 We will add statistics to see which of these works best.
433 2. Modified WsmProcessor EnumerationContext class since the name conflicts
434 with the pull Operation EnumerationContext class.
435 3. Created a new intermediate level of CIMMessage, the CIMPullResponseData
436 Message so that all of the open and Pull responses can use common code
437 since there are only two variables (endOfSequence and EnumerationContext)
438 and they are common across all the open and pull responses.
439 4. Removed a number of diagnostics.
440 5. Cleaned up the Dispcatcher so that the open and pulls have common response
441 code and that code can be used from the dispatcher handle functions and
442 karl 1.2 the provider response functions.
443
444 31 March 2014 - Checkin
445 1. Fixed issues in OOP processing of pull operations, in particular
446 issues with cimxml output format when processed through the
447 *InternalXmlEncoder functions.
448 2. Clean up some of the internalXml functionality
449 3. Found issues causing timeout with a particular provider. The issue
450 is that the dispatcher and monitor end up using the same thread so the
451 condition variable in the dispatcher thread stops the monitor. Turned
452 off the conditionVariable in getCache for the moment which means that
453 we get number of responses for open... with 0 objects before the
454 providers can begin to respond. This is only for test.
455 4. Added some statistics for enumerations and display the statistics
456 when we close the server (same as cache statistics)
457
458 12 March 2014 - Mergeout and Mergein
459 1. Mergeout to head of tree for this date and mergein for patch update
460 to bug 9676
461 2. Extensions to pullop tests program and tests.
462 3. Added some diagnostics in looking for OOP issue.
463 karl 1.2 4. Removed a number of diagnostics messages and cleaned up code in
464 dispatcher to simplify pull operation processing.
465
466 15 December 2013
467 1. Mergeout and mergein up to 15 December 2013
468 2. Clean up issues from tests documented in bug 9676 last week.
469 3. Clean up some code in dispatcher
470 4. Remove the filter function from ResponseStressc++Provider.
471
472 21 November 2013
473 1. Mergeout from head of tree to 21 November 2013.
474
475 18 November 2013
476 1. Cleanup of a bunch of minor errors and completion of all of the code for
477 the openQueryInstances except for the PullInstances in Dispatcher and
478 the aggregator function.
479 2. OpenqueryInstances added to cimcli.
480
481 13 October 2013 CVS branch update.
482 1. Integrated bug 9786 into the branch. Note that we need to test the
483 generated statistics.
484 karl 1.2 2. Mergeout executed to update to head of tree as of 8:00 am 13 October 2013.
485 3. Cleaned up several errors in OOP processing. Note that there is at least
486 one issue left when we to a pull on ManagedElement in at least one of the
487 namespaces.
488 4. Cleaned up some of the outstanding diagnostic code
489 5. Generally passes all tests except for one test of pullop where it is trying
490 to pull enum instances CIM_ManagedElement from a particular namespace.
491
492 NOTE: I did not make comments here for changes in October despite the fact
493 that I did 2 mergouts, number of fixes, and a mergein.
494
495 30 September 2013 - CVS Update
496 Mergeout head of tree up to 29 September 2013.
497
498 29 September 2013. CVS update.
499 1. Modified calls to statisticalData.cpp to a) directly call with request
500 type, b) incorporate the open, pull, etc. messages. However, since these
501 are not part of the CIM class, we must do something special with them.
502 See bug 9785 for full solution to this issue.
503 2. Corrected OOP interface to enable new flag to indicate internal operations
504 and set host, etc.
505 karl 1.2 3. Add code to CQLOperationsDispatcher and WQLOperationDispatcher to clean
506 up CIMResponseDataCounter after filtering.
507 4. Modified ProviderAgent to set Host info for some pull operations.
508 5. Added new flag to CIMBinMsgSerializer and Deserializer.
509
510 17 September 2013 CVS update (Actually two different updates over 3 days)
511 1. Clean up some issues in CIMMessage.h and CIMMessage.cpp
512 2. Extend OpenExecQuery to WQL and CQL processors but return not complete
513 3. Remove memory leak in EnumerationContext and EnumerationContextTable
514 handling.
515 4. Created template functions for much of the pull operations.
516 5. Reversed order of queryLanguage and query (and changed names to match
517 execQuery) in client and server. Note that these are the execQuery
518 WQL and CQL filters and NOT FQL filters.
519 6. Some code cleanup in dispatcher
520 7. Today, passes all tests in pullop but issue in alltests. For some reason
521 not finding CIMObjectManager instance. Also, leaves enumeration contexts
522 if client terminates since cleanup thread not operating.
523 8. XML from OOP not correctly processed.
524
525 14 September 2013 CVS update
526 karl 1.2 Merged out up to 25 August. Cleaned up all operations and standardized code.
527 At this point the non pull operations code is in a set of templates but the
528 pull is not yet.
529 Fixed a significant number of problems so that it appears that the operations
530 except for OpenExecQuery run stably, at least with the pullop test program.
531 Note that there is a problem in that the Interop control provider is not
532 returning its singleton wbemserver object for some reason. Causes a test
533 failure
534
535 Fixed for 16 June CVS Update
536 1. Cleaned up the enumerationContext and Table release functions and tested
537 to confirm that we do not lose memory in either normal sequences or
538 sequences that close early. Cleaned up pullop and added more tests
539 Taged Before: PREAUG25UPDATE and after POSTAUG25UPDATE
540
541 Fixed for 9 June CVS update
542 1. Cleaned up code for OpenQueryInstances. Note that this is incomplete.
543 No support in WQL or CQL Operations
544 2.
545
546 What was fixed for 5 June checkin.
547 karl 1.2 1. Extended ResponseTest MOF for for both CMPI and C++ subclasses
548 2. Fixed issues with pullop.
549 3. Fixed temp issue with CIMResponseData size by putting in mutex. That
550 is not a permanent fix but it gets around issue probably in the control
551 of the move logic that meant counts were off.
552 4. Fixed issues in Dispatcher so that associator code works. Still messy
553 code in the dispatcher.
554 5. Changed name of Enumerationtable.h & cpp to EnumerationContextTable.*
555 6 Changed name of ResponseStressTest module, classes, etc.
556
557 TAG: TASK_PEP317_5JUNE_2013_2
558
559 2 June 2013
560
561 Issues - KS
562
563 - Still way to many TODO and KS comments and KS_TEMPS. Removing bit by bit.
564
565 - Runtime variable connection for the config parameters not installed. That
566 has been made into a separate bug (see bug 9819)
567
568 karl 1.2 5. Issue with the threaded timer. For some reason during tests it
569 eventually calls the timer thread with trash for the parm (which is
570 pointer to the EnumerationTable object). Caught because we do a valid
571 test at beginning of the function.
572
573 6. Still using the templates in CIMOperationRequestDispatcher to simplify
574 the handle... processing.
575
576 7. I think I have a way around the double move of objects in the
577 EnumerationContext so that the outputter will just take a defined number
578 of objects directly from the gathering cache and save the second move.
579
580 8. Not yet passing all tests but getting closer now. The major test that is
581 causing an error today is the execution of a full enumeration with the
582 forceProviders = true. This causes a client timeout sometimes.
583
584
585
586 ===========================================
587
588 OVERVIEW:
589 karl 1.2
590 The operation extensions for pull operations defined in the DMTF specification
591 DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
592 including Client and Server.
593
594 These operations extend the CIM/XML individual operations to operation
595 sequences where the server must maintain state between operations in a
596 sequence and the client must execute multiple operations to get the full
597 set of instances or instance paths.
598
599 The following new CIM/XML operations as defined in DSP0200 are included;
600
601 -OpenEnumerateInstances
602 -openEnumerateInstancePaths
603 -OpenReferenceInstances
604 -OpenReferenceInstancePaths
605 -OpenAssociatiorInstances
606 -OpenAssociatorInstancePaths
607 -OpenQueryInstances
608 -PullInstancesWithPath
609 -PullInstancePaths
610 karl 1.2 -PullInstances
611 -CloseEnumeration
612 -EnumerationCount
613 OpenExecQuery
614
615 The following operations have not been implemented in this version of Pegasus:
616
617 -OpenQueryInstances
618
619 The following limitations on the implementation exist;
620
621 1. The filterQueryLanguage and filterQuery parameters are processed by
622 the Pegasus client but the server returns error if there is any data in
623 either parameter. This work does not include the development of the
624 query language. Note that a separate effort to extend Pegasus to use
625 the DMTF FQL query language is in process.
626
627 2. The input parameter continueOnError is processed correctly by the client
628 but the Pegasus server only provides for false since the server does not
629 include logic to continue processing responses after an error is
630 encountered.
631 karl 1.2 This is consistent with the statement in the specification that use of
632 this functionality is optional and the fact that the DMTF agrees that all
633 of the issues of continuing after errors have not been clarified.
634
635 3. The operation enumerationCount is not processed by the server today since
636 a) really getting the count would be the same cost as the corresponding
637 enumeration, b) the server does not include a history or estimating
638 mechanism for this to date.
639 NOTE: After a through review as part of the development of the next version
640 of CMPI we have concluded that this operation is probably not worth the
641 effort. Since it is optional, Pegasus will only return the unknown status
642 at this point
643
644 Since the concept of sequences of operations linked together (open, pull, close)
645 is a major extension to the original CIM/XML operation concept of completely
646 independent operations several new pieces of functionality are implemented
647 to control interOperationTimeouts, counts of objects to be returned, etc.
648
649 TBD - Review this
650
651 CLIENT
652 karl 1.2
653 The new operations follow the same pattern as the APIs for existing operations
654 in that:
655
656 1. All errors are handled as CIMException and Exception
657
658 2. The means of inputting parameters are the same except that there are
659 significantly more input parameters with the open operations and for the
660 first time operations return parameters as well as objects in the
661 response. Specifically the open and pull operations return values for
662 enumerationContext which is the identity for a pull sequence and
663 endOfSequence which is the marker the server sends in open and pull
664 responses when it has no more objects to send.
665
666 The significant differences include:
667
668 1. Processing of parameters on responses (i.e. the endOfSequence and
669 enumerationContext parameters are returned for open and pull operations).
670
671 2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
672 cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
673 karl 1.2 client api.
674
675 3. The association and reference operations ONLY process instances. They do
676 not include the capability to return classes like reference and associator
677 do and therefore return CIMInstance rather than CIMObject.
678
679 4. Paths are returned in all cases (i.e OpenEnumerateInstances and
680 PullInstancesWithPath where they were not with EnumeratInstances.
681
682 5. The client must maintain state between operations in a sequence (using
683 the enumerationContext parameter).
684
685 TBD- Are there more differences.
686
687
688 SERVER
689
690 The Pegasus server attempts to always deliver the requested number of objects
691 for any open or pull request (the specification allows for the server to
692 deliver less than the requested number of objects and specifically to return
693 zero objects on open). We felt that it was worth any extra cost in processing
694 karl 1.2 to provide the client with exactly what it had requested.
695
696 The pegasus server always closes an enumeration sequence upon receipt of any
697 error from the providers, repository, etc. Therefore the server will reject
698 any request that has continueOnError = true;
699
700 Expansion to allow the continue on error may be added in a future version.
701 In any case, the whole purpose of the continue on error is really to allow
702 input from good providers to be mixed with providers that return errors so
703 that generally this would mean simply changing the logic in the return mechanism
704 to not shutdown when an error is received from any given provider.
705
706 Generally we do not believe that the providers need to do much more in the
707 future to support the continueOnError other than possibly allowing the provider
708 to continue processing after it has received an error.
709
710 PROVIDERS
711
712 This implementation requires NO changes to the existing providers. The
713 provider APIs operate just as they do with the original operations.
714
715 karl 1.2 Because the server processing is different however, there may be some
716 behavior differences primarily because the client now controls the speed of
717 delivery of objects.
718
719 In previous versions of Pegasus, the server attempts to deliver objects as
720 rapidly as then can be put on the network. In the case of HTTP chunked requests
721 they are delivered in chunks of about 100 objects. The primary delay for the
722 providers was the processing of each segment through the server. The server
723 is blocked so that no other segment can proceed through the server until that
724 segment is processed and sent on the network.
725 In the case of non-chunkedresponses, they are completely gathered in the serve
726 and then delivered as one non-chunked response. There were no delays for the
727 providers, just lots of possible memory use in the server.
728
729 The responses from providers (delivered through the deliver(...) interface are
730 gathered into segments of about 100 objects and this group of objects is moved
731 through the server to be delivered to the client.
732
733 However with the inclusion of the pull operations, The segments of objects
734 from the providers are cached in the server response path until the
735 maxObjectCount for that request (open or pull) and that number returned in a
736 karl 1.2 non-chunked response. Thus, if the client is slow to issue pull requests,
737 the providers might be delayed at some point to reduce memory usage in the
738 server (the delay appears as slow response tothe deliver operation).
739
740 In other words, the time to process large sets of responses from the provider
741 now depends on the speed of handling the client.
742
743 It is important to remember in developing providers that the Pegasus server
744 can most efficiently process responses if they are passed from the provider
745 to the server individually or in small arrays of objects rather than the
746 provider gathering very large arrays of objects and sending them to the
747 server.
748
749 NEXT GENERATION PROVIDERS
750 KS_TODO
751
752 CONFIGURATION PARAMETERS
753
754 The server includes several configuration parameters to set limits on the
755 processing of pull operations. All of these configuration parameters are
756 compile time parameters rather than runtime.
757 karl 1.2
758 1. Maximum value of minimum interoperation time. This parameter defines the
759 maximum time allowed between the return of an open or pull response and
760 the receipt of the next pull or a close operation before the server may
761 close the enumeration. The specification allows the server to set a
762 maximum interoperation time and refuse open requests that with requested
763 operationTimeout greater than that time.
764 CIM_ERR_INVALID_OPERATION_TIMEOUT
765
766 This value is set with the Pegasus environment variable
767 PEGASUS_PULL....
768
769 2. Maximum objects returned in a single open or pull operation. The server
770 can set a maximum limit on the number of objects that can be returned in
771 a single open or pull oepration with the maxObjectCount parameter.
772
773 3. Whether the server allows 0 as an interoperation timeout value. The value
774 zero is s special value for the interoperationTimeout in that it tells the
775 server to not timeout any enumeration sequence.
776
777 With this value for interoperationTimeout, the only way to close an
778 karl 1.2 enumeration sequence is to complete all of the pulls or issue the close.
779 If for some reason the sequence is not completed, that enumeration context
780 would remain open indefinitly. Since in Pegasus any open enumeration
781 context uses resources (the context object and any provider resposnes that
782 have not yet been issued in a response) it would appear that most
783 platforms would not want to allow the existence of enumeration contexts
784 that cannot be closed by the server.
785
786 4, maximum consecutive pull requests with 0 maxObjectCount. The use of the
787 pull operation with maxObjectCount set to zero could be used to keep an
788 enumeration context open indefinitly (this tells the server to restart the
789 interoperationTimeout but not send any objects in the response). Therefore the
790 specification allows for the server setting maximum limits on this behavior
791 and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
792 exceeded.
793 Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
794 a non-zero count resets this counter.
795
796 KS-TBD - Is this really logical since we can still block by just issuing
797 lots of zero request and an occansional request for one object.
798
799 karl 1.2 Pegaus sets the value of this limit to 1000 and allows the implementer to
800 modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
801
802 5. Default operationTimeout -
803
804 The default of this parameter is to refuse operat
805
806 In the current release of Pegasus these are all compile time parameters.
807
808
809 NOTES On working with task branch.
810
811 Merge out Process
812
813 To keep our TASK branch in sync with the current head of tree we need
814 to do a regular merge out. the TaskMakefile contains the makefile
815 procedures to do this efficiently. NOTE: Following these procedures is
816 important in that you are merging out new material each time you do
817 the merge out. If you were just to repeatedly merge out, you would be
818 merging previously merged changes a second time causing a real mess.
819
820 karl 1.2 Start with new directory and put TaskMakefile above pegasus (needed so you
821 have this file for the initial operations.
822
823 make -f TaskMakefile branch_merge_out BNAME=PEP317-pullop ## takes a long time
824
825 This checks out current head, merges it into task branch and sets tags
826 for the mergeout. Note that at the end of this step this work is
827 part of the TASK... branch.
828
829 NOW check for conflicts, errors, etc. that resulted from the merge.
830 Look for conflict flags, compare the results (I use linux merge as a
831 good graphic compare tool) and build and test. When you are satisfied
832 that the merge out is clean, you can commit the results to the TASK...
833 branch
834
835 To commit the work to this into Task branch
836
837 make -f mak/TaskMakefile branch_merge_out_commit BNAME=PEP317-pullop
838
839 or manually commit and finish as follows
840
841 karl 1.2 cvs commit
842 make -f mak/TaskMakefile branch_merge_out_finish BNAME=PEP317-pullop
843
844 ## This last step is important since it cleans up temporary tags to prepare
845 you for the next checkout
846
847 COMPARE TASKBRANCH WITH HEAD
848
849 In a new pegasus work space do same as above for merge out.
850
851 make -f TaskMakefile BNAME=PEP317-pullop
852
853 This produces a result which is all of the head merged into the branch.
854 A diff of this is all the new changes to the head of tree that you will
855 include into the merge.
856
|
Return to
readme.pulloperations
CVS log
Up to [Pegasus] / pegasus