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 1.1.2.5 NOTES On working with task branch.
11
12 Merge out
13 I just start with new directory and put TaskMakefile above pegasus
14
15 make -f TaskMakefile BNAME=PEP317-pullop ## takes a long time
16
17 checks out current head, merges it into task branch and sets tags
18 NOW check for conflicts, errors, etc.
19
20 To check this into Task branch
21
22 make -f mak/TaskMakefile branch_merge_out_commit BNAME=PEP317-pullop
23
24 or manually commit and finish as follows
25
26 cvs commit
27
28 make -f mak/TaskMakefile branch_merge_out_finish BNAME=PEP317-pullop
29 ## This last step is important since it cleans up temporary tags.
30
31 karl 1.1.2.5 COMPARE TASKBRANCH WITH HEAD
32
33 In a new pegasus work space do same as for merge out.
34
35 make -f TaskMakefile BNAME=PEP317-pullop
36
37 This produces a result which is all of the head merged into the branch.
38 A diff of this is all the new changes to the head of tree.
39
40
|
41 karl 1.1.2.3
|
42 karl 1.1.2.4 TODO list:
|
43 karl 1.1.2.3 1. Binary operation from OOP. Need to add counter to binary
44 protocol to be able to count objects in response. Generates
45 warnings in things like messageserializer and does not work with
46 OOP right now.
|
47 karl 1.1.2.4 2. OpenExecQuery - Code is incorrect in that it used InstancesWithPath
48 where the spec is instances with no path. Need new function to wrap
49 getInstanceElement(withoutPathElement) in XmlReader. Note that
50 Alternate is to put flag on InstancesWith Path to say no path
51 3. Code for Pull part of OpenQueryInstancesRequest a) should be part of
52 the common CIMOperationRequestDispatcher execCommon code.
53 4. The changes to WQLCIMOperationRequestDispatcher and CQL... for handling
54 pull not completed so we feed the responses back to the EnmerationContext
55 queues
|
56 karl 1.1.2.3 3. Lots of minor TODOs, etc.
|
57 karl 1.1.2.4 4. External runtime variables. Decide this as part of PEP. The variables
58 exist in CIMOperationRequestDispatcher but not in CIMConfig. The primary
59 ones to consider are:
60 a. System maxObjectCount. Setting some maximum size on what a pull
61 client can request (i.e. the maximum size of the maxObjectCount on
62 Open... and pull operations.
63 b. Pull interoperationTimeout (max times between operations). This is
64 the maximum number of seconds on the operationTimeout parameter of the
65 Open operations
66 c. Maximum size of the responseCache before it starts backing up
67 responses to the providers.
|
68 karl 1.1.2.3 5. Decision on EnumerationContext timeout (separate thread or just
|
69 karl 1.1.2.4 checks during other operations). Can we, in fact really keep the
70 enumeration context table and queue under control without monitoring
71 with a separate thread. We must monitor for:
72 a. Client operation that stop requesting (i.e. inter operation time
73 exceeds operationTimeout). Note that if it simply exceeds the time
74 the next operation does the cleanup. The issue is those clients that
75 simply stop and do not either close or go to completion.
76 b. We should protect against providers that no not every finish delivering
77 or take to long between deliveries. This does not exist in Pegasus
78 today
79 6. Clean up code in Dispatcher. The associators code is still real mess
80 and the pull code is in a template. The Pull code is good now but
81 must be duplicated. Look at creating new CIMMessage CIMPullResponseMessage
82 so that we can have common code. Everything is the same except what
83 goes into the CIMResponseData so it is logical to have completely
84 common processing
|
85 karl 1.1.2.3 7. Extension to avoid double move of objects in CIMResponseData (one
86 into enumerationContext queue and second to new cimResponseData for
87 response. Want to avoid second move by extending Open/Pull response
88 messages to include count and CIMResponse data to count objects out
89 of queue when converting (avoids the second move). Big issue here
90 with binary data since need to extend format to count it.
|
91 karl 1.1.2.4 8. Still using templates, etc. in code in the Dispatcher. This is for
92 all of the open operations where there is a lot of duplicate code
93 and the pull operations that are 99% duplicate code (in a single template)
94
95 Fixed for 9 June CVS update
96 1. Cleaned up code for OpenQueryInstances. Note that this is incomplete.
97 No support in WQL or CQL Operations
98 2.
99
100 What was fixed for 5 June checkin.
|
101 karl 1.1.2.3 1. Extended ResponseTest MOF for for both CMPI and C++ subclasses
102 2. Fixed issues with pullop.
103 3. Fixed temp issue with CIMResponseData size by putting in mutex. That
104 is not a permanent fix but it gets around issue probably in the control
105 of the move logic that meant counts were off.
106 4. Fixed issues in Dispatcher so that associator code works. Still messy
107 code in the dispatcher.
108 5. Changed name of Enumerationtable.h & cpp to EnumerationContextTable.*
109 6 Changed name of ResponseStressTest module, classes, etc.
110
111 TAG: TASK_PEP317_5JUNE_2013_2
112
|
113 karl 1.1.2.2 2 June 2013
|
114 karl 1.1.2.1
115 Issues - KS
116 1. have not installed the binary move in CIMResponseData. Please run
117 with OPP off.
118 2. Some problem in the processing so we are getting server crashes.
119 Right no I am guessing that this is in the binaryCodec and am going to
120 expand the test tools to allow testing through the localhost.
121
122 3. Still way to many TODO and KS comments and KS_TEMPS. Removing bit by bit.
123
124 4. Env variable connection for the config parameters not installed.
125
126 5. Issue with the threaded timer. For some reason during tests it
127 eventually calls the timer thread with trash for the parm (which is
128 pointer to the EnumerationTable object). Caught because we do a valid
129 test at beginning of the function.
130
|
131 karl 1.1.2.2 6. Still using the templates in CIMOperationRequestDispatcher to simplify
132 the handle... processing.
133
134 7. I think I have a way around the double move of objects in the
135 EnumerationContext so that the outputter will just take a defined number
136 of objects directly from the gathering cache and save the second move.
137
138 8. Not yet passing all tests but getting closer now.
139
140 9. Created a tag before this commit TASK_PEP317_1JUNE_2013.
141
142 10. Next Tag will be TASK_PEP317_2_JUNE_2013 in the task branch
143
144
|
145 karl 1.1.2.1 ===========================================
146
147 OVERVIEW:
148
149 The operation extensions for pull operations defined in the DMTF specification
150 DSP0200 V 1.4 were implemented in Pegasus effective Pegasus version 2.11
151 including Client and Server.
152
153 These operations extend the CIM/XML individual operations to operation
154 sequences where the server must maintain state between operations in a
155 sequence and the client must execute multiple operations to get the full
156 set of instances or instance paths.
157
158 The following new CIM/XML operations as defined in DSP0200 are included;
159
160 -OpenEnumerateInstances
161 -openEnumerateInstancePaths
162 -OpenReferenceInstances
163 -OpenReferenceInstancePaths
164 -OpenAssociatiorInstances
165 -OpenAssociatorInstancePaths
166 karl 1.1.2.1 -PullInstancesWithPath
167 -PullInstancePaths
168 -CloseEnumeration
169 -EnumerationCount
|
170 karl 1.1.2.2 OpenExecQuery
|
171 karl 1.1.2.1
172 The following operations have not been implemented in this version of Pegasus:
173
174 -OpenQueryInstances
175
176 The following limitations on the implementation exist;
177
178 1. The filterQueryLanguage and filterQuery parameters are processed by
179 the Pegasus client but the server returns error if there is any data in
|
180 karl 1.1.2.2 either parameter. This work does not include the development of the
181 query language. Note that a separate effort to extend Pegasus to use
182 the DMTF FQL query language is in process.
|
183 karl 1.1.2.1
184 2. The input parameter continueOnError is processed correctly by the client
185 but the Pegasus server only provides for false since the server does not
186 include logic to continue processing responses after an error is
187 encountered.
188 This is consistent with the statement in the specification that use of
189 this functionality is optional and the fact that the DMTF agrees that all
190 of the issues of continuing after errors have not been clarified.
191
192 3. The operation enumerationCount is not processed by the server today since
193 a) really getting the count would be the same cost as the corresponding
194 enumeration, b) the server does not include a history or estimating
195 mechanism for this to date.
|
196 karl 1.1.2.2 NOTE: After a through review as part of the development of the next version
197 of CMPI we have concluded that this operation is probably not worth the
198 effort. Since it is optional, Pegasus will only return the unknown status
199 at this point
|
200 karl 1.1.2.1
201 Since the concept of sequences of operations linked together (open, pull, close)
202 is a major extension to the original CIM/XML operation concept of completely
203 independent operations several new pieces of functionality are implemented
204 to control interOperationTimeouts, counts of objects to be returned, etc.
205
206 TBD - Review this
207
208 CLIENT
209
210 The new operations follow the same pattern as the APIs for existing operations
211 in that:
212
213 1. All errors are handled as CIMException and Exception
214
215 2. The means of inputting parameters are the same except that there are
216 significantly more input parameters with the open operations and for the
217 first time operations return parameters as well as objects in the
218 response. Specifically the open and pull operations return values for
219 enumerationContext which is the identity for a pull sequence and
220 endOfSequence which is the marker the server sends in open and pull
221 karl 1.1.2.1 responses when it has no more objects to send.
222
223 The significant differences include:
224
225 1. Processing of parameters on responses (i.e. the endOfSequence and
226 enumerationContext parameters are returned for open and pull operations).
227
228 2. Numeric arguments (Uint32 and Uint64 include the option of NULL in some
229 cases so they are packaged inside classes Uint32Arg and Uint64Arg in the
230 client api.
231
232 3. The association and reference operations ONLY process instances. They do
233 not include the capability to return classes like reference and associator
234 do and therefore return CIMInstance rather than CIMObject.
235
236 4. Paths are returned in all cases (i.e OpenEnumerateInstances and
237 PullInstancesWithPath where they were not with EnumeratInstances.
238
239 5. The client must maintain state between operations in a sequence (using
240 the enumerationContext parameter).
241
242 karl 1.1.2.1 TBD- Are there more differences.
243
244
245 SERVER
246
247 The Pegasus server attempts to always deliver the requested number of objects
248 for any open or pull request (the specification allows for the server to
249 deliver less than the requested number of objects and specifically to return
250 zero objects on open). We felt that it was worth any extra cost in processing
251 to provide the client with exactly what it had requested.
252
253 The pegasus server always closes an enumeration sequence upon receipt of any
254 error from the providers, repository, etc. Therefore the server will reject
|
255 karl 1.1.2.2 any request that has continueOnError = true;
256
257 Expansion to allow the continue on error may be added in a future version.
258 In any case, the whole purpose of the continue on error is really to allow
259 input from good providers to be mixed with providers that return errors so
260 that generally this would mean simply changing the logic in the return mechanism
261 to not shutdown when an error is received from any given provider.
262
263 Generally we do not believe that the providers need to do much more in the
264 future to support the continueOnError other than possibly allowing the provider
265 to continue processing after it has received an error.
|
266 karl 1.1.2.1
267 PROVIDERS
268
269 This implementation requires NO changes to the existing providers. The
270 provider APIs operate just as they do with the original operations.
271
272 Because the server processing is different however, there may be some
273 behavior differences primarily because the client now controls the speed of
274 delivery of objects.
275
276 In previous versions of Pegasus, the server attempts to deliver objects as
277 rapidly as then can be put on the network. In the case of HTTP chunked requests
278 they are delivered in chunks of about 100 objects. The primary delay for the
279 providers was the processing of each segment through the server. The server
280 is blocked so that no other segment can proceed through the server until that
281 segment is processed and sent on the network.
282 In the case of non-chunkedresponses, they are completely gathered in the serve
283 and then delivered as one non-chunked response. There were no delays for the
284 providers, just lots of possible memory use in the server.
285
286 The responses from providers (delivered through the deliver(...) interface are
287 karl 1.1.2.1 gathered into segments of about 100 objects and this group of objects is moved
288 through the server to be delivered to the client.
289
290 However with the inclusion of the pull operations, The segments of objects
291 from the providers are cached in the server response path until the
292 maxObjectCount for that request (open or pull) and that number returned in a
293 non-chunked response. Thus, if the client is slow to issue pull requests,
294 the providers might be delayed at some point to reduce memory usage in the
295 server (the delay appears as slow response tothe deliver operation).
296
297 In other words, the time to process large sets of responses from the provider
298 now depends on the speed of handling the client.
299
300 It is important to remember in developing providers that the Pegasus server
301 can most efficiently process responses if they are passed from the provider
302 to the server individually or in small arrays of objects rather than the
303 provider gathering very large arrays of objects and sending them to the
304 server.
305
|
306 karl 1.1.2.2 NEXT GENERATION PROVIDERS
307 KS_TODO
308
|
309 karl 1.1.2.1 CONFIGURATION PARAMETERS
310
311 The server includes several configuration parameters to set limits on the
312 processing of pull operations. All of these configuration parameters are
313 compile time parameters rather than runtime.
314
315 1. Maximum value of minimum interoperation time. This parameter defines the
316 maximum time allowed between the return of an open or pull response and
317 the receipt of the next pull or a close operation before the server may
318 close the enumeration. The specification allows the server to set a
319 maximum interoperation time and refuse open requests that with requested
320 operationTimeout greater than that time.
321 CIM_ERR_INVALID_OPERATION_TIMEOUT
322
323 This value is set with the Pegasus environment variable
324 PEGASUS_PULL....
325
326 2. Maximum objects returned in a single open or pull operation. The server
327 can set a maximum limit on the number of objects that can be returned in
328 a single open or pull oepration with the maxObjectCount parameter.
329
330 karl 1.1.2.1 3. Whether the server allows 0 as an interoperation timeout value. The value
331 zero is s special value for the interoperationTimeout in that it tells the
332 server to not timeout any enumeration sequence.
333
334 With this value for interoperationTimeout, the only way to close an
335 enumeration sequence is to complete all of the pulls or issue the close.
336 If for some reason the sequence is not completed, that enumeration context
337 would remain open indefinitly. Since in Pegasus any open enumeration
338 context uses resources (the context object and any provider resposnes that
339 have not yet been issued in a response) it would appear that most
340 platforms would not want to allow the existence of enumeration contexts
341 that cannot be closed by the server.
342
343 4, maximum consecutive pull requests with 0 maxObjectCount. The use of the
344 pull operation with maxObjectCount set to zero could be used to keep an
345 enumeration context open indefinitly (this tells the server to restart the
346 interoperationTimeout but not send any objects in the response). Therefore the
347 specification allows for the server setting maximum limits on this behavior
348 and returning the error CIM_ERR_SERVER_LIMITS_EXCEEDED if this limit is
349 exceeded.
350 Note that this is maximum CONSECUTIVE pulls so that issuing a pull with
351 karl 1.1.2.1 a non-zero count resets this counter.
352
353 KS-TBD - Is this really logical since we can still block by just issuing
354 lots of zero request and an occansional request for one object.
355
356 Pegaus sets the value of this limit to 1000 and allows the implementer to
357 modify it with the PEGASUS_MAXIMUM_ZERO_OBJECTCOUNT environment variable.
358
359 5. Default operationTimeout -
360
361 The default of this parameter is to refuse operat
362
363 In the current release of Pegasus these are all compile time parameters.
|