Hi Quentin, Sounds like there maybe a problem with the indexes on the RDF_QUAD table, so you should.
1. Perform integrity check on the database with the command "backup '/dev/null';" 2. Check the count of indexes on the RDF_QUAD table which should all be the same: select count(*) from RDF_QUAD table option (index RDF_QUAD) select count(*) from RDF_QUAD table option (index RDF_QUAD_POGS) select count(*) from RDF_QUAD table option (index RDF_QUAD_SP) If you can provide a lcopy of the DB file it can be checked. ... Best Regards Hugh Williams Professional Services OpenLink Software, Inc. // http://www.openlinksw.com/ Weblog -- http://www.openlinksw.com/blogs/ LinkedIn -- http://www.linkedin.com/company/openlink-software/ Twitter -- http://twitter.com/OpenLink Google+ -- http://plus.google.com/100570109519069333827/ Facebook -- http://www.facebook.com/OpenLinkSoftware Universal Data Access, Integration, and Management Technology Providers On 21 Feb 2014, at 04:12, Quentin <quent...@clearbluewater.com.au> wrote: > Hi Hugh, > > I hadn't, I'd forgotten the name of the audit functions. I ran the > backup->null test described under database recovery and that gave no > indication of a problem. > > Running the audit now as below, it returned '0' and rerunning the query > still failed to return triples with the SPO pattern included. > select DB.DBA.RDF_AUDIT_METADATA(1,'*'); > > However....this has put the database into an interesting state. Previously, > this graph held triples and could be read by some queries. Now it does and > doesn't, depending on how I ask. I ran some of these queries immediately > before the audit and got the expected results. > --- > sparql > prefix o: <http://myOntology/> > > select * > from <http://myGraph/> > { > ?S ?P ?O . > } > --- > Nothing is returned. > --- > sparql > prefix o: <http://myOntology/> > > select * > from <http://myGraph/> > { > ?S a o:myClass . > } > --- > Nothing is returned. > --- > sparql > prefix o: <http://myOntology/> > select ?G ?S > where { GRAPH ?G { > ?S a o:myClass . > }}; > --- > Nothing is returned. > -- > sparql > select distinct ?G > where { GRAPH ?G { > ?S ?P ?O . > }}; > --- > This lists every graph including the graph in question, suggesting that at > least one triple does in fact exist in it. > Conductor->LinkedData->Graphs also lists the graph. > > I'm not sure what to do with the DB now. > > I can provide a login if you're interested in a closer look at it. > > I did copy this database and transaction log to a clean build of develop7 as > of a day ago and that had the same problem though it's unknown whether it > would still be capable of getting into that state. Without knowing how it > got into this state, it's not practical to retest in any particular release > as the intermittence ensures there'll always be a more recent release to test > against when it does occur. > > > > On 21 February 2014 14:37, Hugh Williams <hwilli...@openlinksw.com> wrote: > Hi Quentin, > > Have you tried running the DB.DBA.RDF_AUDIT_METADATA() function which can > check for and repair RDF metadata corruption, as detailed at: > > http://docs.openlinksw.com/virtuoso/fn_rdf_audit_metadata.html > > Best Regards > Hugh Williams > Professional Services > OpenLink Software, Inc. // http://www.openlinksw.com/ > Weblog -- http://www.openlinksw.com/blogs/ > LinkedIn -- http://www.linkedin.com/company/openlink-software/ > Twitter -- http://twitter.com/OpenLink > Google+ -- http://plus.google.com/100570109519069333827/ > Facebook -- http://www.facebook.com/OpenLinkSoftware > Universal Data Access, Integration, and Management Technology Providers > > On 20 Feb 2014, at 05:08, Quentin <quent...@clearbluewater.com.au> wrote: > > > Actually, the explain plan was from a pattern in an earlier testcase that > > also failed. The one listed has this explain plan. > > REPORT > > VARCHAR > > { > > RDF_QUAD_POGS 8 rows(s_10_2_t0.S) > > P = ##type , O = #/APIConsumer G = #/ > > > > Precode: > > 0: S := Call __id2in (s_10_2_t0.S) > > 5: BReturn 0 > > RDF_QUAD_GS unq 0.64 rows () > > inlined G = #/ , S = s_10_2_t0.S > > RDF_QUAD_SP 10 rows(s_10_2_t1.P) > > inlined S = k_s_10_2_t0.S > > RDF_QUAD_POGS 0.3 rows(s_10_2_t1.O) > > inlined P = k_s_10_2_t1.P S = k_s_10_2_t0.S , G = #/ > > > > After code: > > 0: P := Call __id2in (s_10_2_t1.P) > > 5: O := Call __ro2sq (s_10_2_t1.O) > > 10: BReturn 0 > > Select (S, P, O) > > } > > No. of rows in result: 20 > > > > > > On 20 February 2014 16:05, Quentin <quent...@clearbluewater.com.au> wrote: > > This appears to be an intermittent error that pops up every few weeks, > > restarting the DB usually resolves it though not this time. > > > > The query that fails is of the form: > > > > sparql > > prefix o: <http://myOntology/> > > > > select * > > from <http://myGraph/> > > { > > > > ?S a o:myClass . > > #?S ?P ?O . > > } > > ==== > > It works until I remove that comment, correctly returning three bindings > > for ?S. Once that comment is removed, zero bindings are returned. If I > > comment the class line and just return SPO, I can find the ?S in the > > triples that matches the class binding. > > > > The explain plan is this: > > REPORT > > VARCHAR > > { > > RDF_QUAD_POGS 8 rows(s_11_2_t0.S) > > P = ##type , O = #/MyClass G = #/ > > > > Precode: > > 0: S := Call __id2in (s_11_2_t0.S) > > 5: BReturn 0 > > RDF_QUAD_GS unq 0.5 rows () > > inlined G = #/ , S = s_11_2_t0.S > > RDF_QUAD_SP 8.3 rows(s_11_2_t1.P) > > inlined S = k_s_11_2_t0.S > > RDF_QUAD_POGS 0.28 rows(s_11_2_t1.O) > > inlined P = k_s_11_2_t1.P S = k_s_11_2_t0.S , G = #/ > > > > After code: > > 0: P := Call __id2in (s_11_2_t1.P) > > 5: O := Call __ro2sq (s_11_2_t1.O) > > 10: BReturn 0 > > Select (S, P, O) > > } > > No. of rows in result: 20 > > > > Debug log under trace_on shows nothing unusual. > > > > The instance is: > > Version 7.0.1-dev.3207-pthreads as of Feb 3 2014 > > Compiled for Linux (x86_64-unknown-linux-gnu) > > > > I've quarantined the running instance of the DB and it consistently fails > > as above. This behaviour began just a few hours ago, queries of this sort > > performed perfectly this morning and then began failing abruptly with no > > obvious reason why. > > > > I'm not sure it's related but when I cloned the running instance and > > started the DB on the clone, the log includes this error: > > 05:01:06 ERROR: Error executing a server init statement : 22023: SR528: > > Uninitialized property qmfOkForAnySqlvalue in JSO instance > > <http://www.openlinksw.com/virtrdf-data-formats#default-iid> of type > > <http://www.openlinksw.com/schemas/virtrdf#QuadMapFormat> -- > > DB.DBA.RDF_QUAD_FT_UPGRADE () > > > > > > -- > > Quentin | Clear Blue Water Pty Ltd > > quent...@clearbluewater.com.au > > > > > > > > -- > > Quentin | Clear Blue Water Pty Ltd > > quent...@clearbluewater.com.au > > ------------------------------------------------------------------------------ > > Managing the Performance of Cloud-Based Applications > > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > > Read the Whitepaper. > > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk_______________________________________________ > > Virtuoso-users mailing list > > Virtuoso-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/virtuoso-users > > > > > -- > Quentin | Clear Blue Water Pty Ltd > quent...@clearbluewater.com.au > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk_______________________________________________ > Virtuoso-users mailing list > Virtuoso-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/virtuoso-users ------------------------------------------------------------------------------ Flow-based real-time traffic analytics software. Cisco certified tool. Monitor traffic, SLAs, QoS, Medianet, WAAS etc. with NetFlow Analyzer Customize your own dashboards, set traffic alerts and generate reports. Network behavioral analysis & security monitoring. All-in-one tool. http://pubads.g.doubleclick.net/gampad/clk?id=126839071&iu=/4140/ostg.clktrk _______________________________________________ Virtuoso-users mailing list Virtuoso-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/virtuoso-users