RE: Issue regarding CIMServerDiscovery.lookup() method

Hi,

This is to follow up on the issue related to CIMServerDiscovery.lookup() throwing en exception.

From your explanation, we understand that the latest versions of Open Pegasus has changed with respect to how the exception is printed in a more understandablwe way. However, has any improvement been made on the behavior of the API? ie, in case there is an invalid character in one of the URLs, can that URL be ignored so that the disocvery process can proceed with finding other hosts in the subnet? Or, in other words, instead of returning an empty URL list, can the code return all valid URLs?

Thanks,
Hitha

From: V, Hitha
Sent: Wednesday, June 29, 2011 4:20 PM
To: Marek Szermutzky
Cc: [email protected]; Pandurengan, Ponselvarani; CK, Syamlal
Subject: RE: Issue regarding CIMServerDiscovery.lookup() method

Our application uses OP 2.6.

Regards,
Hitha V

From: Marek Szermutzky [mailto:[email protected]]
Sent: Monday, June 27, 2011 4:12 PM
To: V, Hitha
Cc: [email protected]; Pandurengan, Ponselvarani; CK, Syamlal
Subject: Re: Issue regarding CIMServerDiscovery.lookup() method

Hi.


In OpenPegasus 2.11 we introduced an improved version of the exception you see. It prints data up to the "bad" point and hex coded some of the following bytes.
Since this is not the case on your side I assume that you currently use an earlier version of OP. The fix was brought back to OP 2.10.1 which is not released yet though.

What you could do to identify the URL is to put the patch from Bug#8813 onto your OP source and go from there. Please also note the two defines BADUTF8_MAX_CLEAR_CHAR and BADUTF8_MAX_CHAR_TO_HEX which allow you to steer how many clear(=clean=okay) characters are printed and how many of the "bad" characters are encoded into hex chars. For your specific case I recommend to set #define BADUTF8_MAX_CLEAR_CHAR to at least 4000 (so you get the entire clear text string).



Mit freundlichen Grüßen / Kind regards
Marek Szermutzky

Software Engineer / OpenPegasus Maintainer (PMC) and z/OS PlatformRep.
IBM Systems &Technology Group, Systems Software Development / z/OS Capacity Management and Support
-------------------------------------------------------------------------------------------------------------------------------------------
IBM Deutschland
Schoenaicher Str. 220
71032 Boeblingen
Phone: +49-7031-16-5182
E-Mail: [email protected]
-------------------------------------------------------------------------------------------------------------------------------------------
IBM Deutschland Research & Development GmbH / Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Dirk Wittkopp
Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294



From:        "V, Hitha" <[email protected]>
To:        "[email protected]" <[email protected]>
Cc:        "CK, Syamlal" <[email protected]>, "Pandurengan, Ponselvarani" <[email protected]>
Date:        22.06.2011 13:48
Subject:        Issue regarding CIMServerDiscovery.lookup() method
________________________________



Hi,

Our application uses Pegasus SLP APIs to discover the CIMOMs in a network. This method works correctly when invoked from a Linux machine. However, on some Windows machines it has been seen that the method throws the exception “The byte sequence starting at index 3936 is not valid UTF-8 encoding” and subsequently fails to return the CIMOM list.

When I run “slptool” from Linux (I don’t have a Windows version of it) and look at the output URL list, I don’t see any invalid characters.

Has anyone faced a similar issue anytime? If so, what is the work around? How can we identify the URL which contains the invalid character, so that we can isolate it from the network and proceed with testing?

Regards,
Hitha V