[OAI-implementers] error reporting
Irina Dijour
Irina.Dijour@exlibris.co.il
Mon, 24 Jun 2002 16:08:48 +0200
>How about for when the database is down, or other low-level
>problems?
I meant that in the situation when we don't have error codes for low-level
errors
Michael solution #2 is the best solution for my specific problem with
datestamp.
In general I would prefer special error code(s) for low-level errors.
>I for one would like to use something like:
>
><error code="ServerDown">Service unavailable. Retry in one
>hour</error>
>
>with a 503 code, but obviously that's not in the protocol.
Do you mean that your response will contain both error code "ServiceDown"
and HTTP error code 503?
Is it possible?
Thanks,
Irina
-----Original Message-----
From: deridder [mailto:deridder@cs.utk.edu]
Sent: Monday, June 24, 2002 2:54 PM
To: Irina Dijour
Cc: 'Michael L. Nelson'; oai-implementers@oaisrv.nsdl.cornell.edu
Subject: RE: [OAI-implementers] error reporting
How about for when the database is down, or other low-level
problems?
I for one would like to use something like:
<error code="ServerDown">Service unavailable. Retry in one
hour</error>
with a 503 code, but obviously that's not in the protocol.
--j.
On Sun, 23 Jun 2002, Irina Dijour
wrote:
> Hi Michael,
>
> I think that if we don't have error codes for low-level errors your
solution
> #2 (some lower level bound)
> is the most suitable.
>
> Thanks,
> Irina
>
> -----Original Message-----
> From: Michael L. Nelson [mailto:mln@ils.unc.edu]
> Sent: Sunday, June 23, 2002 4:43 PM
> To: oai-implementers@oaisrv.nsdl.cornell.edu
> Subject: RE: [OAI-implementers] error reporting
>
>
> On Sun, 23 Jun 2002, Irina Dijour wrote:
>
> > >What error condition do you wish to report that is not covered
> > >by the existing error codes? I did not find any when writing
> > >my server.
> > For example: One record in my database is wrong and I can not obtain
> > datestamp information for this record.
> > I could say "cannotDisseminateFormat" but in truth I can disseminate
> format,
> > I just cannot get datestamp.
>
> interesting... all of the errors defined in OAI-PMH are about OAI-based
> errors / exception conditions. there is no provision for lower-level
> errors, such as problems in the database.
>
> I think the lack of a datestamp might be enough to say that this record
> does not exist in an OAI-context.
>
> from section 2.5:
>
> A record is identified unambigiously by the combination of the unique
> identifier of the item from which the record is available, the
> metadataPrefix identifying the metadata format of the record, and the
> datestamp of the record.
>
>
> so if you don't have 1 of {id,metadataPrefix,datestamp}, I don't think you
> have a record. I would either:
>
> 1. ignore this record for OAI purposes
>
> or
>
> 2. make up a datestamp for the record. it should not be hard to define
> some lower bound for the record; use whatever you put in earliestDatestamp
> in your Identify response.
>
> regards,
>
> Michael
>
> >
> > BR
> > Irina
> >
> > -----Original Message-----
> > From: Simeon Warner [mailto:simeon@cs.cornell.edu]
> > Sent: Thursday, June 20, 2002 11:16 PM
> > To: Irina Dijour
> > Cc: 'oai-implementers@oaisrv.nsdl.cornell.edu'
> > Subject: Re: [OAI-implementers] error reporting
> >
> >
> >
> >
> > What error condition do you wish to report that is not covered
> > by the existing error codes? I did not find any when writing
> > my server.
> >
> > My feeling is that 503 is not to be used for errors in the
> > same sense as <error> replies. 503 is for cases when the
> > service is unavailable -- it may be used for flow control
> > or perhaps temporary outages such as might occur with large
> > updates.
> >
> > Cheers,
> > Simeon.
> >
> >
> > On Thu, 20 Jun 2002, Irina Dijour wrote:
> > > I have a small suggestion concerning error reporting.
> > >
> > > We have several error codes in OAI 2.0 protocol ("badArgument",
> > > "badResumptionToke" etc).
> > > If I understand correctly, if none of these error codes fits for an
> error
> > > that should be reported,
> > > the OAI data provider should return HTTP Error code "503 Service
> > > unavailable".
> > > 1. If the OAI data provider uses HTTP Error code 503 it cannot specify
> > what
> > > exactly has happened.
> > > And sometimes it's useful.
> > > 2. There are situations when the OAI Data provider wants to report a
> > > specific error for a specific record.
> > > I think that sending HTTP Error code 503 in this case is too
> general.
> > >
> > > Therefore I propose to add additional error "unspecifiedError" (as
it's
> > done
> > > in Z39.50 protocol).
> > > In the "error" element content OAI data provider may specify the
error.
> > >
> > >
> > > Irina
> > >
> > > -----------------------------------
> > > Irina Dijour
> > > Software Development
> > > Ex Libris ltd.
> > > Malha Technology Park
> > > Jerusalem 91481, Israel
> > > Tel: +972-(0)2-6798-222 ext. 142
> > > Fax: +972-(0)2-6798-634
> > > Irina.Dijour@exlibris.co.il
> > >
> > > _______________________________________________
> > > OAI-implementers mailing list
> > > OAI-implementers@oaisrv.nsdl.cornell.edu
> > > http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
> > >
> > _______________________________________________
> > OAI-implementers mailing list
> > OAI-implementers@oaisrv.nsdl.cornell.edu
> > http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
> >
>
> ---
> Michael L. Nelson
> NASA Langley Research Center m.l.nelson@larc.nasa.gov
> MS 158, Hampton, VA 23681 http://www.ils.unc.edu/~mln/
> +1 757 864 8511 +1 757 864 8342 (f)
>
> _______________________________________________
> OAI-implementers mailing list
> OAI-implementers@oaisrv.nsdl.cornell.edu
> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
> _______________________________________________
> OAI-implementers mailing list
> OAI-implementers@oaisrv.nsdl.cornell.edu
> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
>
***********************************************************
PGPKey: http://www.cs.utk.edu/~deridder/jd-pgp.txt
***********************************************************