> -----Original Message-----
> From: Raymond K. Camden [mailto:[EMAIL PROTECTED]]
> Sent: Monday, August 07, 2000 1:07 PM
> To: [EMAIL PROTECTED]
> Subject: RE: Verity Collections in a Cluster- Allaire's recommendation
> to us
>
> > However, following the procedure below works for getting existing
> > collections moved to a shared location, but as soon as you
> try to re-index
> > them once they're on the shared server, we get the same
> errors as before
> > whenever CFINDEX is called. This means the procedure will
> not work for
> > Spectra sites at all (e.g. cfa_metadataindexall fails).
>
> In theory, once you have created the collections on the other
> machines and
> then updated them to point to the shared box, you should never again
> re-index. You should only update/delete and optimize. This is
> true even when
> you don't use one machine for Verity. Check out the
> indexupdate and delete
> tags (both for metadata and type searching).
I'm not sure I understand you. Spectra tags such as
<cfa_metadataindexupdate.cfm> all call <cfindex>, and it is at this point
that the updating fails when the collections are on a remote machine,
whereas it's ok if they're local. The problem seems to occur whatever action
is called.
E.g.
------------------------------------------------------------------
Error Diagnostic Information
Error occurred in tag CFINDEX
Internal Error: Key Delete, Type=File, VdkCollectionSubmit failed
Error Code: -23
The error occurred while processing an element with a general identifier of
(CFINDEX), occupying document position (91:7) to (91:83) in the template
file C:\PROGRAM
FILES\ALLAIRE\SPECTRA\CUSTOMTAGS\SYSTEM\SYSTEM\TIER0\METADATA\CFA_METADATAIN
DEXUPDATE.CFM.
------------------------------------------------------------------
And line 91 of cfa_metadataindexupdate is:
<CFINDEX ACTION="delete" COLLECTION="cfa_metadata_keywords"
KEY="#objectID#">
The only solutions to this dilema that I can think of are 1) To make the
shared machine one of the CF servers, then make sure all admin requests
(i.e. that alter the metadata or type collections) go through that machine.
But this completely defeats the purpose of having a cluster. Or 2) Copy the
c:\blahblah\collections folder from one machine to the shared server every
time somebody updates an object. Which is also a horrific kludge and
probably wouldn't work anyway.
Please tell me there's some other way!
-Tristan.
------------------------------------------------------------------------------
To Unsubscribe visit
http://www.houseoffusion.com/index.cfm?sidebar=lists&body=lists/spectra_talk or send a
message to [EMAIL PROTECTED] with 'unsubscribe' in the body.