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