From: 'Eli Tucker' <he...@ne...> - 2003-03-26 20:30:09
|
Hey guys. Thanks for the info. To clarify, if I get the=20 "INFO Z3950Origin - Target does not support concurrent operations"=20 message, does that mean that performing multiple queries at once on a single provider is not supported, and thus will possibly cause the=20 "WARN Z3950Origin - Unable to locate a search for that REFID. REFID processing at the target may be BROKEN!" message? In other words, in JZKit terms, the following would be unsuccessful: searchTask1.evaluate(-1); searchTask2.evaluate(-1); searchTask3.evaluate(-1); where all searchTasks are to the same Z3950Origin? =20 I decided to do some tests doing exactly that against a server that states they do not support concurrent operations (OCLC), but all tests seem to pass -- I get the expected results back and no broken REFID errors. =20 Is it a fluke that these tests are passing? If not, what does "not supporting concurrent operations" really mean? Thanks, - Eli -----Original Message----- From: Ian Ibbotson [mailto:ian...@k-...]=20 Sent: Wednesday, March 26, 2003 4:09 AM To: LeVan,Ralph Cc: 'Eli Tucker'; Jzj...@li... Subject: RE: [Jzjkit-user] Question about two logging messages. Hiya All... Just as a side note, JZKit uses string based refids to identify which request a response pertains to. It's used primarily when concurrent operations are in force to make sure we can pair request and response messages for different operations when several "SearchTask" objects are active on a single z association. I guess that exception should really log the refid that is causing the problem..... Ian. On Wed, 2003-03-26 at 11:42, LeVan,Ralph wrote: > We don't support concurrent operations, so you'd get that message for sure. >=20 > We've got two, maybe three different servers here. I'm positive that the > one I wrote handles binary Reference ID's and will echo any Reference ID > correctly. I don't have any personal experience with that and the other > servers. >=20 > Ralph >=20 > -----Original Message----- > From: 'Eli Tucker' [mailto:he...@ne...] > Sent: Tuesday, March 25, 2003 3:32 PM > To: 'LeVan,Ralph'; Jzj...@li... > Subject: RE: [Jzjkit-user] Question about two logging messages. >=20 >=20 > Thanks for your response, Ralph. Since you are from OCLC, can you tell > me if OCLC's servers would "croak" and cause this type of error message? >=20 >=20 > - Eli >=20 > -----Original Message----- > From: LeVan,Ralph [mailto:le...@oc...]=20 > Sent: Tuesday, March 25, 2003 12:15 PM > To: 'Eli Tucker'; Jzj...@li... > Subject: RE: [Jzjkit-user] Question about two logging messages. >=20 > > -----Original Message----- > > From: Eli Tucker [mailto:he...@ne...] > > Sent: Tuesday, March 25, 2003 2:47 PM > >=20 > > Does anyone know what the following messages mean? > >=20 > > [Z3950 Search Thread z3950.o.org] INFO Z3950Origin - Target does not > > support concurrent operations >=20 > This refers to a feature in z39.50 to support multiple simultaneous > transactions between a client and a server. For it to work, it needed > to be > negotiated at init time. The message is just telling you that your > client > asked for that feature and the server said "No!". >=20 >=20 > > [Z3950 Search Thread z3950.o.org] WARN Z3950Origin - Unable=20 > > to locate > > a search for that REFID. REFID processing at the target may be BROKEN! >=20 > When your client sends off a search to the server (target), it sticks a > piece of data in every request that the server is required to echo back. > That data is called the reference ID for the request. There are some > services of Z39.50 that require it, but don't worry about that; you > don't > use them. Otherwise, the reference ID is for the convenience of the > client > to help it figure out what request the response corresponds to. (I > know, > the socket you got the response on is even better.) >=20 > What the message is telling you is that the server is not returning your > reference ID correctly. It is getting mangled somehow. Some servers > assumed that the reference ID would always be a string and some clients > (I > don't know about JZKit) put binary data into that field (perfectly > legal) > and the servers croaked. >=20 > I'm guessing that is what this is about. >=20 > Ralph >=20 >=20 >=20 > ------------------------------------------------------- > This SF.net email is sponsored by: > The Definitive IT and Networking Event. Be There! > NetWorld+Interop Las Vegas 2003 -- Register today! > http://ads.sourceforge.net/cgi-bin/redirect.pl?keyn0001en > _______________________________________________ > Jzjkit-user mailing list > Jzj...@li... > https://lists.sourceforge.net/lists/listinfo/jzjkit-user >=20 --=20 Ian Ibbotson (ian...@k-...) Knowledge Integration Ltd Sheffield Science & Technology Parks Cooper Buildings Arundel Street Sheffield S1 2NS http://www.k-int.com |