sblim-devel — All SBLIM-related discussion, including tech support

Re: [Sblim-devel] wseventsink could not receive indication.

[Klaus K. <kk...@su...>]

* Bai, Bin (HPServers-Core-PSC) <bi...@hp...> [Dec 24. 2013 09:22]:
> 
> It looks like the communication between openwsman and wseventsink encounter some problem. Am I right?

Uhm, this could be a selinux problem on Redhat. Try to run with
selinux disabled.

> 
> By the way, I found the following defects:
> 1. I set openwsman to use PAM authentication, it requests pam_unix2.so, however, on RHEL6.5 this library doesn't exist. I have to manually create a link to pam_unix.so.

Well, this problem is known but probably not correctly reflected in
the rpm package or the cmake files.
package/openwsman.pam.rh contains a Redhat-conformant pam configuration.

> 2. I have to install sblim-sfcc-devel to let the whole process work. The sblim-sfcc-devel will create libcimcClientXML.so libcimcclient.so libcmpisfcc.so to those ended with ".so.x".  But the developer package not always available to customer. This problem exist on SLES 11 sp 3 too.

This was a packaging error on the SLE 11 SP3 packages which should be
fixed meanwhile.

Klaus
-- 
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg)
Maxfeldstraße 5, 90409 Nürnberg, Germany

[Sblim-devel] method parameter name

[George v. <geo...@gm...>]

Hi,

I have using sblim- development packages for my provider implementation. In
my provider, classes have methods and this method have parameters also.
I have one requirement was function name & function parameter name should
be case sensitive. There is no difficult to checking method name for this.
But I don't know this fo paramer name case.
Please let me know which method is calling for performing this.

Regards,
George

[Sblim-devel] New Release: SFCB 1.4.7

[David H. <hel...@us...>]

Hi Folks,

We end the year with our final release of 2013: SFCB 1.4.7.  There are a 
few patches to fix some memory leaks with long running sfcc clients, and 
the usual handful of bug fixes, mostly for things that would occur only in 
unusual cases.

There are a couple of new sfcb.cfg properties: sslDhParamsFilePath 
supports the use of an external Diffie Hellman (DH) parameters file, and 
sslEcDhCurveName allows the user to set the "curve name" for Elliptic 
Curve Diffie Hellman (ECDH) key generation.  These options can be used 
with openssl cipher suites supporting "ephemeral" Diffie Hellman (DHE) key 
exchange algorithms.  Ephemeral key exchange enhances security by 
decoupling the private key used to generate the session key, from the 
private key associated with the server certificate.  This provides 
"forward secrecy" while maintaining a key exchange algorithm compatible 
with a conventional RSA-based server certificate.  See the .cfg file 
comments for more info.

Perhaps the most fun new feature is SFCB_LABELPROCS.  This will label each 
SFCB process by adding some identifying information to the cmdline so that 
it appears in 'ps' output.  This provides an easy way to identify each 
running SFCB process without having to rely on the "sfcb-ps.sh" or 
"sfcbproc" utilities.  This really enhances visibility and is a great 
troubleshooting tool.

Further, when using the ps command's -T or H options, you can see a 
labeled view with thread-level granularity, which is something you can't 
really do with "sfcb-ps.sh" or "sfcbproc".  (Individual threads are not 
labeled per se; that is, each thread of a process has the same label.  But 
you can easily see which threads belong to which process, and you can see 
the individual thread ids for each.)

For example, before starting sfcb:

$ export SFCB_LABELPROCS=100

Get a view of running sfcb processes:

$ watch 'ps -efT H | grep sfcb[d]'

Thanks to all who contributed to this release, especially Chris who fixed 
up the SFCC issues and Klaus Kampf who made a big contribution to 
LABELPROCS.

You can see the NEWS file for full list of changes.  The new release is 
tagged (SFCB_1_4_7) in git and the tarball is posted to the SourceForge 
download page.

http://sourceforge.net/projects/sblim/files/sblim-sfcb/1.4%20branch%20%28current%29/

Next release I'll be focusing on some long standing shutdown and process 
hang issues.

-Dave H.

SBLIM Team

[Sblim-devel] wseventsink could not receive indication.

[Bai, B. (HPServers-Core-PSC) <bi...@hp...>]

Hello All,

I successfully using wsman/openwsman/sfcb/wseventsink to observe indications on SLES 11 sp3.
Now my work turn on RHEL 6.5
I can use wsman to enumerate instances. And I can use wsman to subscribe indication, however, wseventsink can't output the indications.
Here is the log from openwsman:

Dec 24 16:07:24 [10716] one heartbeat document created for 36ee65fc-ee43-1e43-8002-a0926d2fb5b4
Dec 24 16:07:24 [10716] wse_heartbeat_sender for 36ee65fc-ee43-1e43-8002-a0926d2fb5b4 started
Dec 24 16:07:24 [10716] Endpoint: http://127.0.0.1:80/eventsink
Dec 24 16:07:24 [10716] cl->authentication.verify_peer: 1
Dec 24 16:07:24 [10716] *****set post buf len = 482******
Dec 24 16:07:27 [10716] curl error code: 1.
Dec 24 16:07:27 [10716] cl->response_code: 0.
Dec 24 16:07:27 [10716] cl->last_error code: 3.
Dec 24 16:07:27 [10716] [ wse_notification_sender thread for 36ee65fc-ee43-1e43-8002-a0926d2fb5b4 quit! ]

The wseventsink continue print the something like:
                At Dec 24 16:07:40 received:

<?xml version="1.0" encoding="UTF-8"?>
<s:Envelope xmlns:s="http://www.w3.org/2003/05/soap-envelope" xmlns:wsa="http://schemas.xmlsoap.org/ws/2004/08/addressing" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <s:Header>
    <wsa:To>http://127.0.0.1:80/eventsink</wsa:To>
    <wsa:Action xs:mustUnderstand="true">http://schemas.xmlsoap.org/wbmem/wsman/1/wsman/Heartbeat</wsa:Action>
    <wsa:MessageID>uuid:3cf8cf63-ee43-1e43-801d-a0926d2fb5b4</wsa:MessageID>
 </s:Header>
  <s:Body/>
</s:Envelope>

It looks like the communication between openwsman and wseventsink encounter some problem. Am I right?

By the way, I found the following defects:
1. I set openwsman to use PAM authentication, it requests pam_unix2.so, however, on RHEL6.5 this library doesn't exist. I have to manually create a link to pam_unix.so.
2. I have to install sblim-sfcc-devel to let the whole process work. The sblim-sfcc-devel will create libcimcClientXML.so libcimcclient.so libcmpisfcc.so to those ended with ".so.x".  But the developer package not always available to customer. This problem exist on SLES 11 sp 3 too.



Thanks and Regards,
Bai, Bin.

Re: [Sblim-devel] regarding profile registration profile

[David H. <hel...@us...>]

Hi George,

At first I thought you were asking about CIM_RegisteredProfile 
specifically.  But now I see, your question is a general one?  You are 
expecting to get the base class as a result of an ei query instead of the 
subclass?

If your provider (defines and) creates a subclass, the subclass is the 
correct response for the ei, ein, gi, etc.

If you are trying to get your provider to create instances of the base 
class directly... I'm not sure if that can be done in a 
standards-compliant way.  I can't think of any examples where that is 
done.  Perhaps other folks on the mailing list have.

Is there a reason you would like to create instances of the base class 
directly instead of a subclass?  As long as you get all the instances of 
the subclass when you query an ei, etc. on the base class -- which should 
always be the case -- does it matter?

Dave

David E. Heller
IBM Linux Technology Center
hel...@us...




George varghese <geo...@gm...> 
12/16/2013 02:23 AM

To
David Heller/Allentown/IBM@IBMUS, 
cc

Subject
Re: [Sblim-devel] regarding profile registration profile






Hi,

When you get free time, please reply. I am waiting for your response.

Regards,
George


On Sat, Nov 30, 2013 at 6:31 PM, George varghese <geo...@gm...> 
wrote:
Hi,

I have one question,

Please see the mof file for CIM_SoftwareIdentity.

class GK_SoftwareIdentity : CIM_SoftwareIdentity
{
};

I expecting a GC or a EI operation on this class to return the base class 
CIM_SoftwareIdentity instead of GK_SoftwareIdentity?

Here ,I have filled some properties values in provider by progrmatically. 
So, I can't declare a static class for this.

Is it possible?

Please advise me.

Regards,
George


On Tue, Nov 26, 2013 at 2:50 PM, George varghese <geo...@gm...> 
wrote:
Dear David,

Thank you. It's very helpful for me. 

Hi George, 
Sorry for the delay.  Apologies also that I may not be understanding your 
question: 

You have registered a new class "GK_RegProfile" but you are expecting a GC 
or a EI operation on this class to return the base class 
CIM_RegisteredProfile instead of GK_RegProfile? 
    yes. That's my requirement. 
 I think returning the result of GK_RegProfile is expected here.
  Ok. 

Or perhaps your intent is create instances of the base 
CIM_RegisteredProfile class directly?  There is a test XML in SFCB xmltest 
directory that will do this, for example: see 
"zcreateInstance_RegProfile.xml".  Also you can create static instances of 
CIM_RegisteredProfile by defining them in a MOF file.  For example, you 
can force one the default instances of SFCB_RegisteredProfile to appear 
instead as instance of CIM_RegisteredProfile by making an edit like this 
to the mof file, then running "make postinstall" to recompile the mof: 
$ vi /var/lib/sfcb/stage/mofs/root/interop/10_interop.mof 
 ... 
 class SFCB_RegisteredProfile : CIM_RegisteredProfile 
  { 
  }; 
-instance of SFCB_RegisteredProfile 
+instance of CIM_RegisteredProfile 
 { 
   InstanceID = "CIM:SFCB_PR"; 
   RegisteredOrganization = 2; 
   RegisteredName = "Profile Registration"; 
   RegisteredVersion = "1.0.0"; 
   AdvertiseTypes = 3; 
 }; 
yes. I have removed the class for this and added above mentioned (instance 
of CIM_RegisteredProfile) things to mof file. then run sfcprepos -f .. and 
try to enumerate class, CIM_RegisteredProfile.  It's working. Thank you.. 
Thank you very much.

You wish your provider to do this?  To create instances of the base class 
CIM_RegisteredProfile rather than some subclass? 

Yes. I think that for implementing profile registration profile, no need 
to create subclass & implementation in c file. But I have one doubt.  I 
wish to change the existing implementation of 
CIM_SoftwareInstallationService class instance like this. But in that 
case, I want to implement InstallFromURI method also?.. In that case is it 
possible? .. how?


Please advise me..

Regards,
George


On Tue, Nov 26, 2013 at 2:27 AM, David Heller <hel...@us...> wrote:
Hi George, 

Sorry for the delay.  Apologies also that I may not be understanding your 
question: 

You have registered a new class "GK_RegProfile" but you are expecting a GC 
or a EI operation on this class to return the base class 
CIM_RegisteredProfile instead of GK_RegProfile?  I think returning the 
result of GK_RegProfile is expected here. 

Or perhaps your intent is create instances of the base 
CIM_RegisteredProfile class directly?  There is a test XML in SFCB xmltest 
directory that will do this, for example: see 
"zcreateInstance_RegProfile.xml".  Also you can create static instances of 
CIM_RegisteredProfile by defining them in a MOF file.  For example, you 
can force one the default instances of SFCB_RegisteredProfile to appear 
instead as instance of CIM_RegisteredProfile by making an edit like this 
to the mof file, then running "make postinstall" to recompile the mof: 

$ vi /var/lib/sfcb/stage/mofs/root/interop/10_interop.mof 
 ... 
 class SFCB_RegisteredProfile : CIM_RegisteredProfile 
  { 
  }; 

-instance of SFCB_RegisteredProfile 
+instance of CIM_RegisteredProfile 
 { 
   InstanceID = "CIM:SFCB_PR"; 
   RegisteredOrganization = 2; 
   RegisteredName = "Profile Registration"; 
   RegisteredVersion = "1.0.0"; 
   AdvertiseTypes = 3; 
 }; 

You wish your provider to do this?  To create instances of the base class 
CIM_RegisteredProfile rather than some subclass? 

Thanks, 
Dave 


David E. Heller
IBM Linux Technology Center
hel...@us...




George varghese <geo...@gm...> 
11/13/2013 05:08 AM 


To
sbl...@li..., 
cc

Subject
Re: [Sblim-devel] regarding profile registration profile








Hi, 

Anybody please reply to me, I am waiting for your response? 

Regards, 
George 


On Mon, Nov 11, 2013 at 2:27 PM, George varghese <geo...@gm...> 
wrote: 
Hi, 

I had implemented a provider for profile registration using 
sblim-development packages. I have some doubts regarding it , For 
implementing it, I have been created a mof file, please see the content of 
the mof file, 

class GK_RegProfile : CIM_RegisteredProfile 
{ 
}; 

After implementation, testing, I got the following response, 
<n1:GK_RegProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>Base</n1:InstanceID>
                        
<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>
                        
<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1::GK_RegProfile>

But I expected following response,

<n1:CIM_RegisteredProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>base</n1:InstanceID>
                        
<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>
                        
<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1:CIM_RegisteredProfile> 

For getting this kind of response, what can I do, or above response is 
correct?. 

For implementing this profile, is it need a new class for this?  
 Can't we use CIM_RegisteredProfile itself?.. if it posible how? 

Please help me. 

Regards, 
George 








On Tue, Jul 16, 2013 at 10:04 PM, Tyrel Datwyler <
ty...@li...> wrote: 
On 07/16/2013 08:05 AM, George varghese wrote:
>
> I want to register these profile as per the standard of profile
> registration profile. How to achieve it using sfcb. How to create
> instances of CIM classes.. which classes they are.
>

The sblim-gather package implements the DMTF's DSP1053 Base Metric
Profile. You could look at its provider implementation for a reference.

-Tyrel


------------------------------------------------------------------------------
DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps
OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access
Free app hosting. Or install the open source package on any LAMP server.
Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!
http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk
_______________________________________________
Sblim-devel mailing list
Sbl...@li...
https://lists.sourceforge.net/lists/listinfo/sblim-devel

[Sblim-devel] New Release: SBLIM Java CIM Client 2.2.5

[Dave B. <de...@us...>]

Today the SBLIM project has released the JSR48 CIM Client for Java 
version 2.2.5.   It is fully compliant with the final DSP0201 2.4 
(Representation of CIM in XML) specification and with the updated final 
JSR48 1.0.0 specification (documentation changes only). It contains two 
new features:

  * Two new Java properties - sblim.wbem.sslClientProtocol and 
sblim.wbem.sslListenerProtocol - can be used to set the protocol used 
for SSLContext.getInstance(String protocol) by a client or listener. (#2647)
  * One new Java property - 
sblim.wbem.cimxmlParser.allowEmptyLocalNameSpacePath - can be used to 
have the client allow LOCALNAMESPACEPATH elements without NAMESPACE 
child elements in incoming responses. (#2711)

Version 2.2.5 also contains five changes that might affect existing 
applications:

  * Three internal APIs - CIMXMLParserImpl.parseNAMESPACE, 
parseClassName and parseCLASSNAME - now throw the CIMXMLParseException. 
(#2670)
  * The internal API CIMXMLBuilderImpl.createMETHODCALL now only has 
three parameters, pParamtype is not needed. (#2689)
  * Two internal APIs - Node.getParamType and CIMObjectFactory.getType - 
now throw the SAXException. (#2693)
     One internal API - PropertyNode.childValueNodeParsed - now throws 
the SAXException. (#2700)
     One internal API - PropertyArrayNode.childValueNodeParsed - now 
throws the SAXException. (#2701)

The complete list of changes follows:

2717 Update detailed release history HTML for 2.2.5
2716 Sync up javax.* javadoc with JSR48 1.0.0 Final V
2719 TCK: CIM APIs should not generate NullPointerException
2594 CR28: Support CIMErrorDescription HTTP field
2718 Bad CIMStatusCode generates NumberFormatException
2538 CR14: Support new CORRELATOR element
2715 Add VALUE.NULL support
2714 Add detailed CIM-XML parser test based on DSP0201 2.4
2713 Enforce loose validation of CIM-XML documents
2712 SimpleReqNode allows any CIM element as child
2709 Lower the level of the EOF message to FINE
2711 LOCALNAMESPACEPATH allows 0 NAMESPACE children
2710 parseVALUEOBJECTWITH(LOCAL)PATH ignores (LOCAL)CLASSPATH child
2708 CIMNode quietly ignores DECLARATION child
2707 INSTANCENAME ignores KEYVALUE and VALUE.REFERENCE children
2706 Bad PARAMETER.REFARRAY ARRAYSIZE generates NumberFormatException
2705 PARAMETER.ARRAY does not require TYPE attribute
2704 PARAMETER does not require TYPE attribute
2703 MethodNode should not require TYPE attribute
2702 Bad PROPERTY.ARRAY ARRAYSIZE generates NumberFormatException
2701 PROPERTY.ARRAY does not require TYPE attribute
2700 PROPERTY does not require TYPE attribute
2699 parseQUALIFIER does not require TYPE attribute
2697 (I)MethodResponseNode allows ERROR with PARAMVALUE
2696 parseIRETURNVALUE ignores VALUE and VALUE.ARRAY
2695 parseMETHODCALL allows LOCALCLASSPATH and LOCALINSTANCEPATH
2694 NAME attribute not required by DOM parser (part 2)
2693 ReturnValueNode allows invalid PARAMTYPE attribute
2691 RETURNVALUE should not require PARAMTYPE attribute
2537 Add new data types for PARAMVALUE
2690 Remove RESPONSEDESTINATION support
2689 createMETHODCALL should not add PARAMTYPE attribute
2688 parseMETHODCALL looks for CIMName attribute instead of NAME
2687 ExpParamValueNode allows VALUE, (I)METHODRESPONSE children
2686 parseEXPPARAMVALUE allows 2+ children, prohibits 0
2685 Element.getAttribute returns empty string if no attribute
2684 parseEXPMETHODRESPONSE has several issues
2683 KEYVALUE VALUETYPE optional, "string" default
2682 (I)MethodCallNode allows no LOCAL*PATH
2681 parseQUALIFIERDECLARATION does not require TYPE attribute
2680 IPARAMVALUE parsing broken on DOM/SAX
2679 parseIMETHODCALL requires one IPARAMVALUE child element
2678 parseMULTI___ allows one SIMPLE___ child element
2677 ObjectPathNode allows all child nodes
2676 parseMULTI(EXP)REQ looking for wrong child elements
2675 CIMXMLParseException messages should contain element name
2674 Null pointer exception in CIMDateTime(String)
2673 NameSpaceNode does not need testCompletness()
2672 Remove SIMPLEREQACK support
2671 Potential null pointer exception in parseERROR
2670 NAME attribute not required by DOM parser
2669 Potential null pointer exception in parseMESSAGE
2668 Potential null pointer exception in parseCIM
2666 CR12: Remove ENUMERATIONCONTEXT

Re: [Sblim-devel] lose some indications if i set the destination of indications as a local file

[David H. <hel...@us...>]

Hi Rusty,

I'm not sure whether to be surprised by this or not!  In my testing with a 
heavy indication flood from SFCB's TestIndicationProvider, this did not 
happen.  And with some reservations I'll say it should not happen.  But I 
need to understand what you saw a bit better to say for sure.

Indication delivery is a multithreaded endeavor.  When a provider makes a 
deliverIndication() upcall, multiple threads, up to 
indicationDeliveryThreadLimit, may be spawned.  You will see these as 
additional threads of interopProvider.  When the ThreadLimit is reached, 
the deliverIndication() upcall will block and the provider will be unable 
to issue a new deliverIndication() until a thread becomes available. 
(Unless of course, your provider attempts another deliverIndication() 
through a new thread of its own; but that call would block as well). 
Assuming the provider handles this properly, no indications should be 
dropped here.

If the indicationDeliveryThreadTimeout is set, interopProvider will give 
up on on creating a new thread after the allotted time.  In this case 
indications *will* be dropped, and you would see a syslog error (from 
InteropProviderInvokeMethod) as a result.  Please let me know if see this 
error.  But this is only when ThreadLimit and ThreadTimeout is set to some 
non-default value.

In cases where the ThreadLimit is not reached, the delivery calls will be 
queued within this interopProvider thread pool, and the threads should all 
complete normally as the backlog of requests is cleared.  The delivery 
thread "completes" by doing it's own invokeMethod() upcall, in turn, to 
indCIMXMLHandler.  (So the interopProvider thread pool only grows to begin 
with when indCIMXMLHandler is not able to process the deliveries fast 
enough.)  However, this invokeMethod() upcall to indCIMXMLHandler is 
single-threaded.  Or to be more precise: only one invokeMethod() upcall is 
permitted at a time.  This means that indCIMXMLHandler should only have 
one thread processing the delivery at a time.

It is this indCIMXMLHandler thread that actually does the delivery.  (Some 
of the code is in indCIMXMLExport but you will see it as a thread of 
indCIMXMLHandler).  When the URL is http or https, it delivers via curl; 
when the URL is file:// it writes to the file.  Because there is only one 
thread of execution here, all the file writes should happen sequentially. 
Even though it is using buffered I/O (i.e. fprintf), the fclose() should 
flush the buffer on each call.  So no indications *should* be dropped 
here, either.

I did some testing by adding a long sleep in exportIndication() in 
indCIMXMLExport, to force it to block, to see if multiple indCIMXMLHandler 
threads would be spawned.  They were not.  What I saw was, the 
interopProvider thread pool filled, and then drained normally once the 
load was removed, and everything behaved as I described above.  (I did 
find one bug in indicationDeliveryThreadTimeout, which I fixed with 
[sfcb#92]  :-)  But I only ever saw one delivery thread of 
indCIMXMLHandler.

Without that sleep, just allowing the file to be written as fast as 
possible, I also saw no problems under heavy load.  I did see the 
interopProvider thread pool grow as I described above, but all indications 
were successfully delivered to the file.  I saw no missing 
SequenceNumbers.

If you want to investigate further please tell me:
1. What version of SFCB are you running?
2. Are you setting any non-default values for 
indicationDeliveryThreadLimit, indicationDeliveryThreadTimeout?
3. Do you see the "dropped" message I mentioned above?
4. Is your provider attempting multiple deliverIndication() calls, or are 
you flooding from multiple indication providers?
5. Is your file:// URL pointing to a local mount, or are you trying to 
write across NFS or something like that?
6. What do you mean by "missed"?  Are you seeing missing sequence numbers 
in the file?  Or some overwritten entries?

I should mention also that I assume you are *not* using Reliable 
Indications.  I don't think RI would help with what you are describing, 
since they only pertain to situations where the delivery failed, such as a 
bad return code from curl.  We would not get a bad rc from the file write, 
unless the fopen() call failed.  Some I'm guessing you are seeing no 
actual failure from exportIndication().

Final notes:
- You can turn on "Indication Provider" tracing (-t 0x00200) and that may 
shed some light.
- If you are interested I'll explain how to examine the running thread 
pools as I described above..

Thanks,
Dave

David E. Heller
IBM Linux Technology Center
hel...@us...




jianbo peng <rus...@gm...> 
12/05/2013 04:58 AM

To
sblim-devel <sbl...@li...>, 
cc

Subject
[Sblim-devel] lose some indications if i set the destination of 
indications as a local file






Hi,
When I set the subscription of indications, I set the indication 
destination as a local file. If sfcb delivers indications too quickly, 
some indications will be missed in the file. But if I set the destination 
as the format, ip + port, I can get all the indications. Do you have any 
idea why this happens? or any solutions to solve this? Thank you.
------------------------------------------------------------------------------
Sponsored by Intel(R) XDK 
Develop, test and display web and hybrid apps with a single code base.
Download it for free now!
http://pubads.g.doubleclick.net/gampad/clk?id=111408631&iu=/4140/ostg.clktrk
_______________________________________________
Sblim-devel mailing list
Sbl...@li...
https://lists.sourceforge.net/lists/listinfo/sblim-devel

[Sblim-devel] lose some indications if i set the destination of indications as a local file

[jianbo p. <rus...@gm...>]

Hi,
When I set the subscription of indications, I set the indication
destination as a local file. If sfcb delivers indications too quickly, some
indications will be missed in the file. But if I set the destination as the
format, ip + port, I can get all the indications. Do you have any idea why
this happens? or any solutions to solve this? Thank you.

Re: [Sblim-devel] regarding profile registration profile

[David H. <hel...@us...>]

Hi George,

Sorry for the delay.  Apologies also that I may not be understanding your 
question:

You have registered a new class "GK_RegProfile" but you are expecting a GC 
or a EI operation on this class to return the base class 
CIM_RegisteredProfile instead of GK_RegProfile?  I think returning the 
result of GK_RegProfile is expected here.

Or perhaps your intent is create instances of the base 
CIM_RegisteredProfile class directly?  There is a test XML in SFCB xmltest 
directory that will do this, for example: see 
"zcreateInstance_RegProfile.xml".  Also you can create static instances of 
CIM_RegisteredProfile by defining them in a MOF file.  For example, you 
can force one the default instances of SFCB_RegisteredProfile to appear 
instead as instance of CIM_RegisteredProfile by making an edit like this 
to the mof file, then running "make postinstall" to recompile the mof:

$ vi /var/lib/sfcb/stage/mofs/root/interop/10_interop.mof
 ...
 class SFCB_RegisteredProfile : CIM_RegisteredProfile
  {
  };

-instance of SFCB_RegisteredProfile
+instance of CIM_RegisteredProfile
 {
   InstanceID = "CIM:SFCB_PR";
   RegisteredOrganization = 2;
   RegisteredName = "Profile Registration";
   RegisteredVersion = "1.0.0";
   AdvertiseTypes = 3;
 };

You wish your provider to do this?  To create instances of the base class 
CIM_RegisteredProfile rather than some subclass?

Thanks,
Dave


David E. Heller
IBM Linux Technology Center
hel...@us...




George varghese <geo...@gm...> 
11/13/2013 05:08 AM

To
sbl...@li..., 
cc

Subject
Re: [Sblim-devel] regarding profile registration profile






Hi,

Anybody please reply to me, I am waiting for your response?

Regards,
George


On Mon, Nov 11, 2013 at 2:27 PM, George varghese <geo...@gm...> 
wrote:
Hi,

I had implemented a provider for profile registration using 
sblim-development packages. I have some doubts regarding it , For 
implementing it, I have been created a mof file, please see the content of 
the mof file,

class GK_RegProfile : CIM_RegisteredProfile
{
};

After implementation, testing, I got the following response,
<n1:GK_RegProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>Base</n1:InstanceID>
                        
<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>
                        
<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1::GK_RegProfile>

But I expected following response,

<n1:CIM_RegisteredProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>base</n1:InstanceID>
                        
<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>
                        
<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1:CIM_RegisteredProfile>

For getting this kind of response, what can I do, or above response is 
correct?.

For implementing this profile, is it need a new class for this? 
 Can't we use CIM_RegisteredProfile itself?.. if it posible how?

Please help me.

Regards,
George








On Tue, Jul 16, 2013 at 10:04 PM, Tyrel Datwyler <
ty...@li...> wrote:
On 07/16/2013 08:05 AM, George varghese wrote:
>
> I want to register these profile as per the standard of profile
> registration profile. How to achieve it using sfcb. How to create
> instances of CIM classes.. which classes they are.
>

The sblim-gather package implements the DMTF's DSP1053 Base Metric
Profile. You could look at its provider implementation for a reference.

-Tyrel


------------------------------------------------------------------------------
DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps
OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access
Free app hosting. Or install the open source package on any LAMP server.
Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!
http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk
_______________________________________________
Sblim-devel mailing list
Sbl...@li...
https://lists.sourceforge.net/lists/listinfo/sblim-devel

[Sblim-devel] samples for modifyinstance

[George v. <geo...@gm...>]

Hi All,

I had using sblim-development packages for implementing our provider.
I want to implement 'put' action(WSManagement spec) in provider. For this,
I think that implementation comes in modifyinstance.. right?

Please advise for implementing it?
We don't know the how to parse the selctor set properties values.

Please provide a sample file for referring.

Regards,
George

Re: [Sblim-devel] regarding profile registration profile

[David H. <hel...@us...>]

George varghese <geo...@gm...> wrote on 11/13/2013 05:08:22 AM:

> George varghese <geo...@gm...> 
> 11/13/2013 05:08 AM
> 
> To
> 
> sbl...@li..., 
> 
> cc
> 
> Subject
> 
> Re: [Sblim-devel] regarding profile registration profile
> 
> Hi,
> 
> Anybody please reply to me, I am waiting for your response?
> 
> Regards,
> George
>
 
Sorry, I have not had a chance to look at this yet.  I will respond 
shortly. -DaveH


> On Mon, Nov 11, 2013 at 2:27 PM, George varghese <geo...@gm...> 
wrote:
> Hi,
> 
> I had implemented a provider for profile registration using sblim-
> development packages. I have some doubts regarding it , For 
> implementing it, I have been created a mof file, please see the 
> content of the mof file,
> 
> class GK_RegProfile : CIM_RegisteredProfile
> {
> };
> 
> After implementation, testing, I got the following response,
> <n1:GK_RegProfile>
>                         <n1:Caption xsi:nil="true"/>
>                         <n1:Description xsi:nil="true"/>
>                         <n1:ElementName xsi:nil="true"/>
>                         <n1:Generation xsi:nil="true"/>
>                         <n1:InstanceID>Base</n1:InstanceID>
>                         <n1:OtherRegisteredOrganization>***</
> n1:OtherRegisteredOrganization>
>                         <n1:OtherSpecificationType xsi:nil="true"/>
>                         <n1:RegisteredName>***</n1:RegisteredName>
>                         <n1:RegisteredOrganization>*</
> n1:RegisteredOrganization>
>                         <n1:RegisteredVersion>**</n1:RegisteredVersion>
>                         <n1:SpecificationType xsi:nil="true"/>
>                     </n1::GK_RegProfile>
> 
> But I expected following response,
> 
> <n1:CIM_RegisteredProfile>
>                         <n1:Caption xsi:nil="true"/>
>                         <n1:Description xsi:nil="true"/>
>                         <n1:ElementName xsi:nil="true"/>
>                         <n1:Generation xsi:nil="true"/>
>                         <n1:InstanceID>base</n1:InstanceID>
>                         <n1:OtherRegisteredOrganization>***</
> n1:OtherRegisteredOrganization>
>                         <n1:OtherSpecificationType xsi:nil="true"/>
>                         <n1:RegisteredName>***</n1:RegisteredName>
>                         <n1:RegisteredOrganization>*</
> n1:RegisteredOrganization>
>                         <n1:RegisteredVersion>**</n1:RegisteredVersion>
>                         <n1:SpecificationType xsi:nil="true"/>
>                     </n1:CIM_RegisteredProfile>
> 
> For getting this kind of response, what can I do, or above response 
> is correct?.
> 
> For implementing this profile, is it need a new class for this? 
>  Can't we use CIM_RegisteredProfile itself?.. if it posible how?
> 
> Please help me.
> 
> Regards,
> George
> 
> 

> On Tue, Jul 16, 2013 at 10:04 PM, Tyrel Datwyler 
<ty...@li...
> > wrote:
> On 07/16/2013 08:05 AM, George varghese wrote:
> >
> > I want to register these profile as per the standard of profile
> > registration profile. How to achieve it using sfcb. How to create
> > instances of CIM classes.. which classes they are.
> >

> The sblim-gather package implements the DMTF's DSP1053 Base Metric
> Profile. You could look at its provider implementation for a reference.
> 
> -Tyrel

> 
> 
------------------------------------------------------------------------------
> DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps
> OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access
> Free app hosting. Or install the open source package on any LAMP server.
> Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!
> 
http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk

> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel

Re: [Sblim-devel] invalid key data type 12 c

[Klaus K. <kk...@su...>]

* George varghese <geo...@gm...> [Nov 13. 2013 11:41]:
> Hi,
> 
> I am trying to implement a provider using sblim development packages. In
> mof file, one property type named real64. In implementation side, I have
> set the value to this property. After registering library with sfcb, I
> can't enumerate tha class,
> 
> please see the error ,
> cimXmlGen.c(317): invalid key data type 12 c
> 

Apparently, real values are not allowed as key values. sfcb only allows
string, datetime, integer, boolean, or reference.


Klaus
-- 
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg)
Maxfeldstraße 5, 90409 Nürnberg, Germany

[Sblim-devel] invalid key data type 12 c

[George v. <geo...@gm...>]

Hi,

I am trying to implement a provider using sblim development packages. In
mof file, one property type named real64. In implementation side, I have
set the value to this property. After registering library with sfcb, I
can't enumerate tha class,

please see the error ,
cimXmlGen.c(317): invalid key data type 12 c

Please help me.

Regards,
George

Re: [Sblim-devel] regarding profile registration profile

[George v. <geo...@gm...>]

Hi,

Anybody please reply to me, I am waiting for your response?

Regards,
George


On Mon, Nov 11, 2013 at 2:27 PM, George varghese <geo...@gm...>wrote:

> Hi,
>
> I had implemented a provider for profile registration using
> sblim-development packages. I have some doubts regarding it , For
> implementing it, I have been created a mof file, please see the content of
> the mof file,
>
> class GK_RegProfile : CIM_RegisteredProfile
> {
> };
>
> After implementation, testing, I got the following response,
> <n1:GK_RegProfile>
>                         <n1:Caption xsi:nil="true"/>
>                         <n1:Description xsi:nil="true"/>
>                         <n1:ElementName xsi:nil="true"/>
>                         <n1:Generation xsi:nil="true"/>
>                         <n1:InstanceID>Base</n1:InstanceID>
>
> <n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
>                         <n1:OtherSpecificationType xsi:nil="true"/>
>                         <n1:RegisteredName>***</n1:RegisteredName>
>
> <n1:RegisteredOrganization>*</n1:RegisteredOrganization>
>                         <n1:RegisteredVersion>**</n1:RegisteredVersion>
>                         <n1:SpecificationType xsi:nil="true"/>
>                     </n1::GK_RegProfile>
>
> But I expected following response,
>
> <n1:CIM_RegisteredProfile>
>                         <n1:Caption xsi:nil="true"/>
>                         <n1:Description xsi:nil="true"/>
>                         <n1:ElementName xsi:nil="true"/>
>                         <n1:Generation xsi:nil="true"/>
>                         <n1:InstanceID>base</n1:InstanceID>
>
> <n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
>                         <n1:OtherSpecificationType xsi:nil="true"/>
>                         <n1:RegisteredName>***</n1:RegisteredName>
>
> <n1:RegisteredOrganization>*</n1:RegisteredOrganization>
>                         <n1:RegisteredVersion>**</n1:RegisteredVersion>
>                         <n1:SpecificationType xsi:nil="true"/>
>                     </n1:CIM_RegisteredProfile>
>
> For getting this kind of response, what can I do, or above response is
> correct?.
>
> For implementing this profile, is it need a new class for this?
>  Can't we use CIM_RegisteredProfile itself?.. if it posible how?
>
> Please help me.
>
> Regards,
> George
>
>
>
>
>
>
>
>
> On Tue, Jul 16, 2013 at 10:04 PM, Tyrel Datwyler <
> ty...@li...> wrote:
>
>> On 07/16/2013 08:05 AM, George varghese wrote:
>> >
>> > I want to register these profile as per the standard of profile
>> > registration profile. How to achieve it using sfcb. How to create
>> > instances of CIM classes.. which classes they are.
>> >
>>
>> The sblim-gather package implements the DMTF's DSP1053 Base Metric
>> Profile. You could look at its provider implementation for a reference.
>>
>> -Tyrel
>>
>>
>

Re: [Sblim-devel] regarding profile registration profile

[George v. <geo...@gm...>]

Hi,

I had implemented a provider for profile registration using
sblim-development packages. I have some doubts regarding it , For
implementing it, I have been created a mof file, please see the content of
the mof file,

class GK_RegProfile : CIM_RegisteredProfile
{
};

After implementation, testing, I got the following response,
<n1:GK_RegProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>Base</n1:InstanceID>

<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>

<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1::GK_RegProfile>

But I expected following response,

<n1:CIM_RegisteredProfile>
                        <n1:Caption xsi:nil="true"/>
                        <n1:Description xsi:nil="true"/>
                        <n1:ElementName xsi:nil="true"/>
                        <n1:Generation xsi:nil="true"/>
                        <n1:InstanceID>base</n1:InstanceID>

<n1:OtherRegisteredOrganization>***</n1:OtherRegisteredOrganization>
                        <n1:OtherSpecificationType xsi:nil="true"/>
                        <n1:RegisteredName>***</n1:RegisteredName>

<n1:RegisteredOrganization>*</n1:RegisteredOrganization>
                        <n1:RegisteredVersion>**</n1:RegisteredVersion>
                        <n1:SpecificationType xsi:nil="true"/>
                    </n1:CIM_RegisteredProfile>

For getting this kind of response, what can I do, or above response is
correct?.

For implementing this profile, is it need a new class for this?
 Can't we use CIM_RegisteredProfile itself?.. if it posible how?

Please help me.

Regards,
George








On Tue, Jul 16, 2013 at 10:04 PM, Tyrel Datwyler
<ty...@li...>wrote:

> On 07/16/2013 08:05 AM, George varghese wrote:
> >
> > I want to register these profile as per the standard of profile
> > registration profile. How to achieve it using sfcb. How to create
> > instances of CIM classes.. which classes they are.
> >
>
> The sblim-gather package implements the DMTF's DSP1053 Base Metric
> Profile. You could look at its provider implementation for a reference.
>
> -Tyrel
>
>

Re: [Sblim-devel] sfcb 1.4.5 - https connection

[David H. <hel...@us...>]

Hi George,

This is actually a bug in wbemcli.  Please see:

https://sourceforge.net/p/sblim/bugs/2629/

and let me know if you have questions following that.

Thanks,
Dave

David E. Heller
IBM Linux Technology Center
hel...@us...


George varghese <geo...@gm...> wrote on 10/22/2013 06:36:57 AM:

> George varghese <geo...@gm...> 
> 10/22/2013 06:36 AM
> 
> To
> 
> sbl...@li..., 
> 
> cc
> 
> Subject
> 
> [Sblim-devel] sfcb 1.4.5 - https connection
> 
> Hi,
> 
> I had configure sfcb using --enable-ssl option. 
> Then I created cert file using as genSslCert.sh script file. the 
> opied the files in /etc/ssl folder.
> 
> ## These options only apply if configured with --enable-ssl
> 
> ## Enable HTTPS. 
> ## Default is false. If HTTPS is configured, default is true.
> enableHttps:    true
> 
> ## The HTTP port that SFCB should listen on for secure connections. 
> ## Default is 5989
> httpsPort:      5989
> 
> ## Filename containing the private key for the server's certificate.The 
file 
> ## must be in PEM format and may not be passphrase-protected. The file 
is
> ## relevant for both client connect and indications sent via https.
> ## For client connect: the file must be present if enableHttps is true.
> ## For indications: the file is required only if the indication receiver
> ## will attempt to verify the sender (sfcb) certificate.
> ## Default is /etc/sfcb/file.pem
> sslKeyFilePath: /etc/ssl/certs/file.pem
> 
> ## Filename containing the server's certificate. Must be in PEM format.
> ## The file is relevant for both client connect and indications sentvia 
https.
> ## For client connect: the file must be present if enableHttps is true.
> ## For indications: the file is required only if the indication receiver
> ## will attempt to verify the sender (sfcb) certificate.
> ## Default is /etc/sfcb/server.pem
> sslCertificateFilePath: /etc/ssl/certs/server.pem
> 
> ## Filename containing list of certificates server accepts.
> ## The file is relevant client connect only.
> ## Default is /etc/sfcb/clist.pem
> sslCertList: /etc/ssl/clist.pem
> 
> ## How SFCB handles client certificate based authentication. 
> ## ignore - do not request a certificate from the client
> ## accept - request a certificate from the client; do not fail if 
> not presented
> ## require - refuse the client connection if the client doesn't present 
a 
> ##           certificate 
> ## Default is ignore
> sslClientCertificate: ignore
> 
> ## How SFCB handles verification of the endpoint certificate when 
sending
> ## an indication via https.
> ## ignore - do not attempt to validate the endpoint certificate
> ## verify - validate the certificate against known CA certs in the 
> trust store;
> ##          do not send the indication if verification fails.
> ## verifyhostname - additionally validate the certificate CN (common 
name)
> ##          against the indication handler Destination address; 
> ##          do not send the indication if verification fails.
> ## Default is ignore (always send the indication)
> sslIndicationReceiverCert: ignore
> 
> ## Location of the trust store. Contains one or more CA certificates.
> ## The file is relevant for both client connect and indications sentvia 
https.
> ## For client connect: if sslClientCertificate is set to "require", 
> ## certificate presented must present valid according to the trust 
store.
> ## For indications: if sslIndicationReceiverCert is set to "verify" or
> ## "verifyhostname", the endpoint's certificate is checked against this 
file.
> ## Default: /etc/sfcb/client.pem
> sslClientTrustStore: /etc/ssl/client.pem
> 
> ## Name of the local library to call for client-certificate based user 
> ## authentication. 
> ## Applicable only if sslClientCertificate is not set to "accept" 
or"require".
> ## Default is sfcCertificateAuthentication
> certificateAuthLib:   sfcCertificateAuthentication
> 
> ## List of SSL ciphers to enable.
> ## Default is "ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH" which disables
> ##      weak ciphers.
> sslCiphers: ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH
> 
> I got error when I request the following command..
> 
> $wbemcli ecn https://localhost/root/cimv2
> 
> *
> * wbemcli: Http Exception: Problem with the SSL CA cert (path? access 
rights?)
> *
> 
> displaying following logs in sfcbd ,
> 
> --- SSL_ERROR_SYSCALL during handshake: EOF occurred: client may have 
aborted
> 
> *** httpAdapter.c:1638 SSL_ERROR_SYSCALL error during SSL handshake -- 
exiting
> 
> How to resolve this issue?
> 
> Regards,
> George
> 
> 
------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most 
from 
> the latest Intel processors and coprocessors. See abstracts and register 
>
> 
http://pubads.g.doubleclick.net/gampad/clk?id=60135991&iu=/4140/ostg.clktrk

> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel

[Sblim-devel] sfcb 1.4.5 - https connection

[George v. <geo...@gm...>]

Hi,

I had configure sfcb using --enable-ssl option.
Then I created cert file using as genSslCert.sh script file. the opied the
files in /etc/ssl folder.

## These options only apply if configured with --enable-ssl

## Enable HTTPS.
## Default is false. If HTTPS is configured, default is true.
enableHttps:    true

## The HTTP port that SFCB should listen on for secure connections.
## Default is 5989
httpsPort:      5989

## Filename containing the private key for the server's certificate. The
file
## must be in PEM format and may not be passphrase-protected. The file is
## relevant for both client connect and indications sent via https.
## For client connect: the file must be present if enableHttps is true.
## For indications: the file is required only if the indication receiver
## will attempt to verify the sender (sfcb) certificate.
## Default is /etc/sfcb/file.pem
sslKeyFilePath: /etc/ssl/certs/file.pem

## Filename containing the server's certificate. Must be in PEM format.
## The file is relevant for both client connect and indications sent via
https.
## For client connect: the file must be present if enableHttps is true.
## For indications: the file is required only if the indication receiver
## will attempt to verify the sender (sfcb) certificate.
## Default is /etc/sfcb/server.pem
sslCertificateFilePath: /etc/ssl/certs/server.pem

## Filename containing list of certificates server accepts.
## The file is relevant client connect only.
## Default is /etc/sfcb/clist.pem
sslCertList: /etc/ssl/clist.pem

## How SFCB handles client certificate based authentication.
## ignore - do not request a certificate from the client
## accept - request a certificate from the client; do not fail if not
presented
## require - refuse the client connection if the client doesn't present a
##           certificate
## Default is ignore
sslClientCertificate: ignore

## How SFCB handles verification of the endpoint certificate when sending
## an indication via https.
## ignore - do not attempt to validate the endpoint certificate
## verify - validate the certificate against known CA certs in the trust
store;
##          do not send the indication if verification fails.
## verifyhostname - additionally validate the certificate CN (common name)
##          against the indication handler Destination address;
##          do not send the indication if verification fails.
## Default is ignore (always send the indication)
sslIndicationReceiverCert: ignore

## Location of the trust store. Contains one or more CA certificates.
## The file is relevant for both client connect and indications sent via
https.
## For client connect: if sslClientCertificate is set to "require",
## certificate presented must present valid according to the trust store.
## For indications: if sslIndicationReceiverCert is set to "verify" or
## "verifyhostname", the endpoint's certificate is checked against this
file.
## Default: /etc/sfcb/client.pem
sslClientTrustStore: /etc/ssl/client.pem

## Name of the local library to call for client-certificate based user
## authentication.
## Applicable only if sslClientCertificate is not set to "accept" or
"require".
## Default is sfcCertificateAuthentication
certificateAuthLib:   sfcCertificateAuthentication

## List of SSL ciphers to enable.
## Default is "ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH" which disables
##      weak ciphers.
sslCiphers: ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH



I got error when I request the following command..

$wbemcli ecn https://localhost/root/cimv2

*
* wbemcli: Http Exception: Problem with the SSL CA cert (path? access
rights?)
*

displaying following logs in sfcbd ,

--- SSL_ERROR_SYSCALL during handshake: EOF occurred: client may have
aborted

*** httpAdapter.c:1638 SSL_ERROR_SYSCALL error during SSL handshake --
exiting

How to resolve this issue?

Regards,
George

Re: [Sblim-devel] Regarding provider impementation

[George v. <geo...@gm...>]

Hi All,

At last, ,my serious issue is resolved. I used to sblim-development package
for generating skelton c code. Then the issue is resolved. Thanks all of
you for responding my mails.

keep in touch.

Regards,
George


On Thu, Sep 5, 2013 at 3:55 PM, Ales Ledvinka <ale...@re...> wrote:

> Hello,
>
> For ARM related patches you might try to check the Fedora ARM project.
> https://fedoraproject.org/wiki/Architectures/ARM
>
>
> ----- Original Message -----
> From: "George varghese" <geo...@gm...>
> To: sbl...@li...
> Sent: Monday, August 19, 2013 8:43:01 AM
> Subject: Re: [Sblim-devel] Regarding provider impementation
>
> Hi,
>
> Please note the system details, STLinux, ARM
>
> In Linux/fedora system its working fine.
>
> Please help me.
>
> Regards,
> George
>
>
>
> On Wed, Aug 14, 2013 at 6:14 PM, George varghese < geo...@gm... >
> wrote:
>
>
>
> Thanks Klaus for giving your valuable time for me.
>
> In Linux/fedora system its working fine.
>
> But I want to work it in STLinux, ARM
>
> did you know about it?
>
> Regards,
> George
>
>
>
>
> On Wed, Aug 14, 2013 at 5:17 PM, Klaus Kaempf < kk...@su... > wrote:
>
>
> * George varghese < geo...@gm... > [Aug 14. 2013 12:05]:
> > Thank you.
> >
> > konkret cmpi version is konkretcmpi-0.9.0.
>
> That should be recent enough.
>
> >
> > Please note that I was working on system is *STLinux, , ARM, Version is
> > 2.6.37.6*
> > * Processor - SPEAr 1340*
>
> Sorry, I don't have access to such a system. Does your provider work
> for you on a x86/x86_64 Linux system ?
>
> Klaus
> --
> SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix
> Imendörffer, HRB 16746 (AG Nürnberg)
> Maxfeldstraße 5, 90409 Nürnberg, Germany
>
>
>
>
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel
>
>
> ------------------------------------------------------------------------------
> Learn the latest--Visual Studio 2012, SharePoint 2013, SQL 2012, more!
> Discover the easy way to master current and previous Microsoft technologies
> and advance your career. Get an incredible 1,500+ hours of step-by-step
> tutorial videos with LearnDevNow. Subscribe today and save!
> http://pubads.g.doubleclick.net/gampad/clk?id=58041391&iu=/4140/ostg.clktrk
> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel
>

[Sblim-devel] New Release: SFCC 2.2.6

[Chris B. <buc...@li...>]

Hi All,

SFCC 2.2.6 has been released. The release tarball is up on SourceForge, 
and the code has been tagged in git.

Not too many changes, but a few worth mentioning:

+ getClass does not expose method information
+ sfcc cannot reconnect with sfcb if sfcb restarts - ReleaseCIMCEnv() 
now works as expected

A complete list of changes is available in the git log.

As always, thanks to those in the community who contributed code to this 
release.


-Chris
SBLIM Team

Re: [Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause in cimXmlOps.y / cimXmlParserProcessed.c

[David H. <hel...@us...>]

Hi Ting,

I just checked in the patch to upstream SFCB v1.3 and 1.4.

Thanks again for your help here!

Dave H.


David E. Heller
IBM Linux Technology Center
hel...@us...




Ting Wang <wan...@gm...> 
10/12/2013 05:28 AM

To
David Heller/Allentown/IBM@IBMUS, 
cc
sbl...@li...
Subject
Re: [Sblim-devel] An issue related with sfcbd InvokeMethod and its root 
cause in cimXmlOps.y / cimXmlParserProcessed.c






Hello David,

One point I forgot to mention is that the sfcbd I tried with is the one 
shipped with VMware ESXi5.1, it is VMware customized based on 1.3.7 
version.

Please see my comments inline <Ting>.

Regards,
Ting


On Sat, Oct 12, 2013 at 7:13 AM, David Heller <hel...@us...> wrote:
Hi Ting, 

I think you are right, the type should not be set to -1 there.  There is 
no other place in the code where we do this; in all other places where the 
type is not applicable or indeterminate, we set type = 0.  That was 
probably the intent here. 

I did a quick test with similar XML and the SFCB TestMethodProvider.  I 
modified the callmethod1.xml file in ./test/TestProviders/test/, 
eliminating the VALUE, and turning the PARAMVALUE into a self-closing tag 
as you have.  And it did hit the same parser code, and set the type as you 
described.  However in my test, the flow never hit the freeParamValue() 
code, since the call to that comes from freeParamValues(), and in this 
case there are no "xtokParamValues" to free.  That's because nothing gets 
added in buildInvokeMethodRequest(), because p->value.value would be null 
after parsing this XML.  So CMAddArg() never gets called. 
 
<Ting> I agree with you: latest sfcbd should not have the same problem 
that I observed, since between 1.3.7 and main stream latest version, 
buildInvokeMethodRequest() has been implemented which protect against null 
p->value.value just as you mentioned. 
 
The freeParamValue() DOES get called when I actually pass a param there, 
such as the case of the unmodified callmethod1.xml file. 

So I question whether freeParamValue() really got called in the case of 
the XML you pasted above.  More likely, the problem is just that SFCB 
passed this bad type value to your provider, and your provider was unable 
to handle it. 

I will go ahead and fix this bug.  Can you confirm that changing to 
$$.type=0 does indeed fix your problem? 

<Ting> I did a double check regarding the "$$.type=0" fix, and confirm 
that fixes the problem I'm facing on 1.3.7 VMware sfcbd.

One caveat: this bug is preventing the code from entering the 
updateMethodParamTypes() function in buildInvokeMethodRequest(). Once I 
make the change it will enter that function, which attempts to determine 
the correct CMPIType by checking how it is defined in the class repository 
(i.e. how it is defined in the .mof file when it was compiled).  In this 
case it should start picking up the correct CMPIType for this parameter 
(which I assume is "reference", as it is in your XML) and it will pass 
THAT type to your provider, instead of the bad value it is passing now. 
 At that point your provider will have to handle the empty PARAMVALUE, 
which I assume it is prepared to do, or you may have a different problem. 

<Ting> Our provider is prepared to accept null parameter from method 
calls, and I guess most provider should such checks.
 
Just for curiosity, is there a reason you are passing an empty PARAMVALUE 
here instead of no PARAMVALUE at all?  Or is this just the XML that was 
generated by your client code when it had no param to pass? 

<Ting> This message is generated by SNIA-CTP smitest1.4.0.35 test suite 
test case: 
org.snia.smi.ctp.recipes.SMI_BlockServices_StoragePool_Creation_Removal
 
Thanks for bringing this forward. 

-Dave H 


David E. Heller
IBM Linux Technology Center
hel...@us...




Ting Wang <wan...@gm...> 
10/09/2013 04:34 AM 


To
sbl...@li..., 
cc

Subject
[Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause 
in cimXmlOps.y / cimXmlParserProcessed.c








Hello,

We found conformance related problem in CIM-XML processing, could someone 
look at the problem and help me conform? Thank you. 

Problem description:

We found that following CIM-XML message does not get any response from 
provider, although log shows that the method get invoked successfully.

"""
POST /cimom HTTP/1.1
Connection: Keep-Alive
Content-Type: text/xml;charset=UTF-8
Date: Wed, 09 Oct 2013 06:08:51 GMT
Authorization: Basic cm9vdDoxMjM0NTY3ODkw
Content-Length: 591
CIMObject: 
root%2Fxyzsd%3AXYZ_VRAID_StoragePool.InstanceID%3D%22primordial_afdasd%3AVRAID_vmhba1%22
Accept-Language: *
Host: ?.???.??.???:5988
User-Agent: WBEM Solutions WBEM Java API3.1.2 07/16/2009 10:59
CIMOperation: MethodCall
CIMProtocolVersion: 1.0
CIMMethod: GetAvailableExtents
Accept: text/xml, application/xml

<?xml version="1.0" encoding="UTF-8"?><CIM CIMVERSION="2.0" 
DTDVERSION="2.0"><MESSAGE ID="2013:9:9:2:1:8:51:15" 
PROTOCOLVERSION="1.0"><SIMPLEREQ><METHODCALL 
NAME="GetAvailableExtents"><LOCALINSTANCEPATH><LOCALNAMESPACEPATH><NAMESPACE 
NAME="root"/><NAMESPACE NAME="xyzsd"/></LOCALNAMESPACEPATH><INSTANCENAME 
CLASSNAME="XYZ_VRAID_StoragePool"><KEYBINDING NAME="InstanceID"><KEYVALUE 
TYPE="string" 
VALUETYPE="string">primordial_afdasd:VRAID_vmhba1</KEYVALUE></KEYBINDING></INSTANCENAME></LOCALINSTANCEPATH><PARAMVALUE 
NAME="Goal" 
PARAMTYPE="reference"/></METHODCALL></SIMPLEREQ></MESSAGE></CIM>
"""

Through debugging, we identified that the following fields in the CIM-XML 
message caused the problem:
<PARAMVALUE NAME="Goal" PARAMTYPE="reference"/>
After we remove this parameter from the CIM-XML message, CIM client can 
get response from sfcbd.

Root cause:

After careful analysis, we found the problem is caused by incorrect 
handling in cimXmlOps.y:

paramValue
    : XTOK_PARAMVALUE ZTOK_PARAMVALUE
    {
       $$.value.value=NULL;
       $$.type=-1; // In case the paramValue is blank, type is assigned to 
0xFFFFFFFF
    }

However in cimXmlParserProcessed.c, the processing depends on "type". If 
we previously hit the blank paramValue, then type will be 0xFFFFFFFF, and 
it will match the first check, drop into freeRefArray() logic. As a 
result, the logic will hit sanity check in free(), no response is sent 
back due to this error.
static void freeParamValue(XtokParamValue *op)
{
  if ((op->type & CMPI_refA) == CMPI_refA) {
    freeRefArray(&op->valueRefArray);
  } else if (op->type & CMPI_ARRAY) {
    freeArray(&op->valueArray);
  } else if ((op->type & CMPI_ref) == CMPI_ref) {
    freeReference(&op->valueRef);
  } else if ((op->type & CMPI_instance) == CMPI_instance
             || (op->type & CMPI_class) == CMPI_class) {
    freeValue(&op->value);
  }
}

Regards,
Ting
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most 
from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
Sblim-devel mailing list
Sbl...@li...
https://lists.sourceforge.net/lists/listinfo/sblim-devel

Re: [Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause in cimXmlOps.y / cimXmlParserProcessed.c

[Ting W. <wan...@gm...>]

Hello David,

One point I forgot to mention is that the sfcbd I tried with is the one
shipped with VMware ESXi5.1, it is VMware customized based on 1.3.7 version.

Please see my comments inline <Ting>.

Regards,
Ting


On Sat, Oct 12, 2013 at 7:13 AM, David Heller <hel...@us...> wrote:

> Hi Ting,
>
> I think you are right, the type should not be set to -1 there.  There is
> no other place in the code where we do this; in all other places where the
> type is not applicable or indeterminate, we set type = 0.  That was
> probably the intent here.
>
> I did a quick test with similar XML and the SFCB TestMethodProvider.  I
> modified the callmethod1.xml file in ./test/TestProviders/test/,
> eliminating the VALUE, and turning the PARAMVALUE into a self-closing tag
> as you have.  And it did hit the same parser code, and set the type as you
> described.  However in my test, the flow never hit the freeParamValue()
> code, since the call to that comes from freeParamValues(), and in this case
> there are no "xtokParamValues" to free.  That's because nothing gets added
> in buildInvokeMethodRequest(), because p->value.value would be null after
> parsing this XML.  So CMAddArg() never gets called.
>

<Ting> I agree with you: latest sfcbd should not have the same problem that
I observed, since between 1.3.7 and main stream latest version,
buildInvokeMethodRequest()
has been implemented which protect against null p->value.value just as you
mentioned.

>

The freeParamValue() DOES get called when I actually pass a param there,
> such as the case of the unmodified callmethod1.xml file.
>
> So I question whether freeParamValue() really got called in the case of
> the XML you pasted above.  More likely, the problem is just that SFCB
> passed this bad type value to your provider, and your provider was unable
> to handle it.
>
> I will go ahead and fix this bug.  Can you confirm that changing to
> $$.type=0 does indeed fix your problem?
>
> <Ting> I did a double check regarding the "$$.type=0" fix, and confirm
that fixes the problem I'm facing on 1.3.7 VMware sfcbd.

One caveat: this bug is preventing the code from entering the
> updateMethodParamTypes() function in buildInvokeMethodRequest(). Once I
> make the change it will enter that function, which attempts to determine
> the correct CMPIType by checking how it is defined in the class repository
> (i.e. how it is defined in the .mof file when it was compiled).  In this
> case it should start picking up the correct CMPIType for this parameter
> (which I assume is "reference", as it is in your XML) and it will pass THAT
> type to your provider, instead of the bad value it is passing now.  At that
> point your provider will have to handle the empty PARAMVALUE, which I
> assume it is prepared to do, or you may have a different problem.
>
> <Ting> Our provider is prepared to accept null parameter from method
calls, and I guess most provider should such checks.


> Just for curiosity, is there a reason you are passing an empty PARAMVALUE
> here instead of no PARAMVALUE at all?  Or is this just the XML that was
> generated by your client code when it had no param to pass?
>
> <Ting> This message is generated by SNIA-CTP smitest1.4.0.35 test suite
test case:
org.snia.smi.ctp.recipes.SMI_BlockServices_StoragePool_Creation_Removal


> Thanks for bringing this forward.
>
> -Dave H
>
>
> David E. Heller
> IBM Linux Technology Center
> hel...@us...
>
>
>
>  *Ting Wang <wan...@gm...>*
>
> 10/09/2013 04:34 AM
>   To
> sbl...@li...,
> cc
>   Subject
> [Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause
> in cimXmlOps.y / cimXmlParserProcessed.c
>
>
>
>
> Hello,
>
> We found conformance related problem in CIM-XML processing, could someone
> look at the problem and help me conform? Thank you.
>
> Problem description:
>
> We found that following CIM-XML message does not get any response from
> provider, although log shows that the method get invoked successfully.
>
> """
> POST /cimom HTTP/1.1
> Connection: Keep-Alive
> Content-Type: text/xml;charset=UTF-8
> Date: Wed, 09 Oct 2013 06:08:51 GMT
> Authorization: Basic cm9vdDoxMjM0NTY3ODkw
> Content-Length: 591
> CIMObject:
> root%2Fxyzsd%3AXYZ_VRAID_StoragePool.InstanceID%3D%22primordial_afdasd%3AVRAID_vmhba1%22
> Accept-Language: *
> Host: ?.???.??.???:5988
> User-Agent: WBEM Solutions WBEM Java API3.1.2 07/16/2009 10:59
> CIMOperation: MethodCall
> CIMProtocolVersion: 1.0
> CIMMethod: GetAvailableExtents
> Accept: text/xml, application/xml
>
> <?xml version="1.0" encoding="UTF-8"?><CIM CIMVERSION="2.0"
> DTDVERSION="2.0"><MESSAGE ID="2013:9:9:2:1:8:51:15"
> PROTOCOLVERSION="1.0"><SIMPLEREQ><METHODCALL
> NAME="GetAvailableExtents"><LOCALINSTANCEPATH><LOCALNAMESPACEPATH><NAMESPACE
> NAME="root"/><NAMESPACE NAME="xyzsd"/></LOCALNAMESPACEPATH><INSTANCENAME
> CLASSNAME="XYZ_VRAID_StoragePool"><KEYBINDING NAME="InstanceID"><KEYVALUE
> TYPE="string"
> VALUETYPE="string">primordial_afdasd:VRAID_vmhba1</KEYVALUE></KEYBINDING></INSTANCENAME></LOCALINSTANCEPATH><PARAMVALUE
> NAME="Goal" PARAMTYPE="reference"/></METHODCALL></SIMPLEREQ></MESSAGE></CIM>
> """
>
> Through debugging, we identified that the following fields in the CIM-XML
> message caused the problem:
> <PARAMVALUE NAME="Goal" PARAMTYPE="reference"/>
> After we remove this parameter from the CIM-XML message, CIM client can
> get response from sfcbd.
>
> Root cause:
>
> After careful analysis, we found the problem is caused by incorrect
> handling in cimXmlOps.y:
>
> paramValue
>     : XTOK_PARAMVALUE ZTOK_PARAMVALUE
>     {
>        $$.value.value=NULL;
>        $$.type=-1; // In case the paramValue is blank, type is assigned to
> 0xFFFFFFFF
>     }
>
> However in cimXmlParserProcessed.c, the processing depends on "type". If
> we previously hit the blank paramValue, then type will be 0xFFFFFFFF, and
> it will match the first check, drop into freeRefArray() logic. As a result,
> the logic will hit sanity check in free(), no response is sent back due to
> this error.
> static void freeParamValue(XtokParamValue *op)
> {
>   if ((op->type & CMPI_refA) == CMPI_refA) {
>     freeRefArray(&op->valueRefArray);
>   } else if (op->type & CMPI_ARRAY) {
>     freeArray(&op->valueArray);
>   } else if ((op->type & CMPI_ref) == CMPI_ref) {
>     freeReference(&op->valueRef);
>   } else if ((op->type & CMPI_instance) == CMPI_instance
>              || (op->type & CMPI_class) == CMPI_class) {
>     freeValue(&op->value);
>   }
> }
>
> Regards,
> Ting
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel
>
>

Re: [Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause in cimXmlOps.y / cimXmlParserProcessed.c

[David H. <hel...@us...>]

Hi Ting,

I think you are right, the type should not be set to -1 there.  There is 
no other place in the code where we do this; in all other places where the 
type is not applicable or indeterminate, we set type = 0.  That was 
probably the intent here.

I did a quick test with similar XML and the SFCB TestMethodProvider.  I 
modified the callmethod1.xml file in ./test/TestProviders/test/, 
eliminating the VALUE, and turning the PARAMVALUE into a self-closing tag 
as you have.  And it did hit the same parser code, and set the type as you 
described.  However in my test, the flow never hit the freeParamValue() 
code, since the call to that comes from freeParamValues(), and in this 
case there are no "xtokParamValues" to free.  That's because nothing gets 
added in buildInvokeMethodRequest(), because p->value.value would be null 
after parsing this XML.  So CMAddArg() never gets called.

The freeParamValue() DOES get called when I actually pass a param there, 
such as the case of the unmodified callmethod1.xml file.

So I question whether freeParamValue() really got called in the case of 
the XML you pasted above.  More likely, the problem is just that SFCB 
passed this bad type value to your provider, and your provider was unable 
to handle it.

I will go ahead and fix this bug.  Can you confirm that changing to 
$$.type=0 does indeed fix your problem?

One caveat: this bug is preventing the code from entering the 
updateMethodParamTypes() function in buildInvokeMethodRequest(). Once I 
make the change it will enter that function, which attempts to determine 
the correct CMPIType by checking how it is defined in the class repository 
(i.e. how it is defined in the .mof file when it was compiled).  In this 
case it should start picking up the correct CMPIType for this parameter 
(which I assume is "reference", as it is in your XML) and it will pass 
THAT type to your provider, instead of the bad value it is passing now. At 
that point your provider will have to handle the empty PARAMVALUE, which I 
assume it is prepared to do, or you may have a different problem.

Just for curiosity, is there a reason you are passing an empty PARAMVALUE 
here instead of no PARAMVALUE at all?  Or is this just the XML that was 
generated by your client code when it had no param to pass?

Thanks for bringing this forward.

-Dave H


David E. Heller
IBM Linux Technology Center
hel...@us...




Ting Wang <wan...@gm...> 
10/09/2013 04:34 AM

To
sbl...@li..., 
cc

Subject
[Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause 
in cimXmlOps.y / cimXmlParserProcessed.c






Hello,

We found conformance related problem in CIM-XML processing, could someone 
look at the problem and help me conform? Thank you.

Problem description:

We found that following CIM-XML message does not get any response from 
provider, although log shows that the method get invoked successfully.

"""
POST /cimom HTTP/1.1
Connection: Keep-Alive
Content-Type: text/xml;charset=UTF-8
Date: Wed, 09 Oct 2013 06:08:51 GMT
Authorization: Basic cm9vdDoxMjM0NTY3ODkw
Content-Length: 591
CIMObject: 
root%2Fxyzsd%3AXYZ_VRAID_StoragePool.InstanceID%3D%22primordial_afdasd%3AVRAID_vmhba1%22
Accept-Language: *
Host: ?.???.??.???:5988
User-Agent: WBEM Solutions WBEM Java API3.1.2 07/16/2009 10:59
CIMOperation: MethodCall
CIMProtocolVersion: 1.0
CIMMethod: GetAvailableExtents
Accept: text/xml, application/xml

<?xml version="1.0" encoding="UTF-8"?><CIM CIMVERSION="2.0" 
DTDVERSION="2.0"><MESSAGE ID="2013:9:9:2:1:8:51:15" 
PROTOCOLVERSION="1.0"><SIMPLEREQ><METHODCALL 
NAME="GetAvailableExtents"><LOCALINSTANCEPATH><LOCALNAMESPACEPATH><NAMESPACE 
NAME="root"/><NAMESPACE NAME="xyzsd"/></LOCALNAMESPACEPATH><INSTANCENAME 
CLASSNAME="XYZ_VRAID_StoragePool"><KEYBINDING NAME="InstanceID"><KEYVALUE 
TYPE="string" 
VALUETYPE="string">primordial_afdasd:VRAID_vmhba1</KEYVALUE></KEYBINDING></INSTANCENAME></LOCALINSTANCEPATH><PARAMVALUE 
NAME="Goal" 
PARAMTYPE="reference"/></METHODCALL></SIMPLEREQ></MESSAGE></CIM>
"""

Through debugging, we identified that the following fields in the CIM-XML 
message caused the problem:
<PARAMVALUE NAME="Goal" PARAMTYPE="reference"/>
After we remove this parameter from the CIM-XML message, CIM client can 
get response from sfcbd.

Root cause:

After careful analysis, we found the problem is caused by incorrect 
handling in cimXmlOps.y:

paramValue
    : XTOK_PARAMVALUE ZTOK_PARAMVALUE
    {
       $$.value.value=NULL;
       $$.type=-1; // In case the paramValue is blank, type is assigned to 
0xFFFFFFFF
    }

However in cimXmlParserProcessed.c, the processing depends on "type". If 
we previously hit the blank paramValue, then type will be 0xFFFFFFFF, and 
it will match the first check, drop into freeRefArray() logic. As a 
result, the logic will hit sanity check in free(), no response is sent 
back due to this error.
static void freeParamValue(XtokParamValue *op)
{
  if ((op->type & CMPI_refA) == CMPI_refA) {
    freeRefArray(&op->valueRefArray);
  } else if (op->type & CMPI_ARRAY) {
    freeArray(&op->valueArray);
  } else if ((op->type & CMPI_ref) == CMPI_ref) {
    freeReference(&op->valueRef);
  } else if ((op->type & CMPI_instance) == CMPI_instance
             || (op->type & CMPI_class) == CMPI_class) {
    freeValue(&op->value);
  }
}

Regards,
Ting
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most 
from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
Sblim-devel mailing list
Sbl...@li...
https://lists.sourceforge.net/lists/listinfo/sblim-devel

[Sblim-devel] An issue related with sfcbd InvokeMethod and its root cause in cimXmlOps.y / cimXmlParserProcessed.c

[Ting W. <wan...@gm...>]

Hello,

We found conformance related problem in CIM-XML processing, could someone
look at the problem and help me conform? Thank you.

Problem description:

We found that following CIM-XML message does not get any response from
provider, although log shows that the method get invoked successfully.

"""
POST /cimom HTTP/1.1
Connection: Keep-Alive
Content-Type: text/xml;charset=UTF-8
Date: Wed, 09 Oct 2013 06:08:51 GMT
Authorization: Basic cm9vdDoxMjM0NTY3ODkw
Content-Length: 591
CIMObject:
root%2Fxyzsd%3AXYZ_VRAID_StoragePool.InstanceID%3D%22primordial_afdasd%3AVRAID_vmhba1%22
Accept-Language: *
Host: ?.???.??.???:5988
User-Agent: WBEM Solutions WBEM Java API3.1.2 07/16/2009 10:59
CIMOperation: MethodCall
CIMProtocolVersion: 1.0
CIMMethod: GetAvailableExtents
Accept: text/xml, application/xml

<?xml version="1.0" encoding="UTF-8"?><CIM CIMVERSION="2.0"
DTDVERSION="2.0"><MESSAGE ID="2013:9:9:2:1:8:51:15"
PROTOCOLVERSION="1.0"><SIMPLEREQ><METHODCALL
NAME="GetAvailableExtents"><LOCALINSTANCEPATH><LOCALNAMESPACEPATH><NAMESPACE
NAME="root"/><NAMESPACE NAME="xyzsd"/></LOCALNAMESPACEPATH><INSTANCENAME
CLASSNAME="XYZ_VRAID_StoragePool"><KEYBINDING
NAME="InstanceID"><KEYVALUETYPE="string"
VALUETYPE="string">primordial_afdasd:VRAID_vmhba1
</KEYVALUE></KEYBINDING></INSTANCENAME></LOCALINSTANCEPATH><PARAMVALUENAME="Goal"
PARAMTYPE="reference"/></METHODCALL></SIMPLEREQ></MESSAGE></CIM>
"""

Through debugging, we identified that the following fields in the CIM-XML
message caused the problem:
<PARAMVALUE NAME="Goal" PARAMTYPE="reference"/>
After we remove this parameter from the CIM-XML message, CIM client can get
response from sfcbd.

Root cause:

After careful analysis, we found the problem is caused by incorrect
handling in cimXmlOps.y:

paramValue
    : XTOK_PARAMVALUE ZTOK_PARAMVALUE
    {
       $$.value.value=NULL;
       $$.type=-1; // In case the paramValue is blank, type is assigned to
0xFFFFFFFF
    }

However in cimXmlParserProcessed.c, the processing depends on "type". If we
previously hit the blank paramValue, then type will be 0xFFFFFFFF, and it
will match the first check, drop into freeRefArray() logic. As a result,
the logic will hit sanity check in free(), no response is sent back due to
this error.
static void freeParamValue(XtokParamValue *op)
{
  if ((op->type & CMPI_refA) == CMPI_refA) {
    freeRefArray(&op->valueRefArray);
  } else if (op->type & CMPI_ARRAY) {
    freeArray(&op->valueArray);
  } else if ((op->type & CMPI_ref) == CMPI_ref) {
    freeReference(&op->valueRef);
  } else if ((op->type & CMPI_instance) == CMPI_instance
             || (op->type & CMPI_class) == CMPI_class) {
    freeValue(&op->value);
  }
}

Regards,
Ting

Re: [Sblim-devel] Regarding provider impementation

[George v. <geo...@gm...>]

Hi,

Issue still exists.
Right now I am trying to generate skeleton c code using sblim development
package(sblim-psg-0.8.1).

I have created a .c file

$sblim-psg-0.8.1/cmpi# ./createCmpiProvider.sh GK_Test cmpiGK_TestProvider
inst

it creates cmpiGK_TestProvider.c file..

Have any tutorial for build & make and register it.

Please help me.

Regards,
George






On Thu, Sep 5, 2013 at 3:55 PM, Ales Ledvinka <ale...@re...> wrote:

> Hello,
>
> For ARM related patches you might try to check the Fedora ARM project.
> https://fedoraproject.org/wiki/Architectures/ARM
>
>
> ----- Original Message -----
> From: "George varghese" <geo...@gm...>
> To: sbl...@li...
> Sent: Monday, August 19, 2013 8:43:01 AM
> Subject: Re: [Sblim-devel] Regarding provider impementation
>
> Hi,
>
> Please note the system details, STLinux, ARM
>
> In Linux/fedora system its working fine.
>
> Please help me.
>
> Regards,
> George
>
>
>
> On Wed, Aug 14, 2013 at 6:14 PM, George varghese < geo...@gm... >
> wrote:
>
>
>
> Thanks Klaus for giving your valuable time for me.
>
> In Linux/fedora system its working fine.
>
> But I want to work it in STLinux, ARM
>
> did you know about it?
>
> Regards,
> George
>
>
>
>
> On Wed, Aug 14, 2013 at 5:17 PM, Klaus Kaempf < kk...@su... > wrote:
>
>
> * George varghese < geo...@gm... > [Aug 14. 2013 12:05]:
> > Thank you.
> >
> > konkret cmpi version is konkretcmpi-0.9.0.
>
> That should be recent enough.
>
> >
> > Please note that I was working on system is *STLinux, , ARM, Version is
> > 2.6.37.6*
> > * Processor - SPEAr 1340*
>
> Sorry, I don't have access to such a system. Does your provider work
> for you on a x86/x86_64 Linux system ?
>
> Klaus
> --
> SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix
> Imendörffer, HRB 16746 (AG Nürnberg)
> Maxfeldstraße 5, 90409 Nürnberg, Germany
>
>
>
>
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel
>
>
> ------------------------------------------------------------------------------
> Learn the latest--Visual Studio 2012, SharePoint 2013, SQL 2012, more!
> Discover the easy way to master current and previous Microsoft technologies
> and advance your career. Get an incredible 1,500+ hours of step-by-step
> tutorial videos with LearnDevNow. Subscribe today and save!
> http://pubads.g.doubleclick.net/gampad/clk?id=58041391&iu=/4140/ostg.clktrk
> _______________________________________________
> Sblim-devel mailing list
> Sbl...@li...
> https://lists.sourceforge.net/lists/listinfo/sblim-devel
>

[Sblim-devel] New Release: SFCB 1.4.6

[David H. <hel...@us...>]

Hi Folks,

We have a new SFCB release today: SFCB 1.4.6.  Now supported is the 
ability to reload SSL context without restarting any SFCB process, which 
is an improvement over the previous design of having to restart the HTTP 
daemon.  This is useful for updating certs or making other SSL 
configuration changes on the fly.  Other new features include a new 
sfcb.cfg property to control the number of instances returned per response 
when HTTP chunking is used.

There is the usual handful of bugfixes, including a few that affect 
long-running localconnect clients.  There were a number of small 
improvements, fixes and code-cleanup contributed by the community this 
time, particularly Klaus and Rusty.  Thanks, guys!

Please see the NEWS file for full list of changes.  The new release is 
tagged (SFCB_1_4_6) in git and the tarball is posted to the SourceForge 
download page.  Enjoy!

http://sourceforge.net/projects/sblim/files/sblim-sfcb/1.4%20branch%20%28current%29/

-Dave H.

SBLIM Team