Hi Hugh,
The dataset is geonames.org's ~120million triple data set. I
converted the quasi-RDF/XML dump they have (URI, RDF/XML. URI, RDF/XML,
URI, RDF/XML ...) and split the triples over 50 turtle files.
I repeated the same experiement again for the third time (with what
I thought was the same protocol I had followed), but the bug did not occur.
I repeated the same experiment again but I remembered that I had
hit Conductor-->Linked Data-->Graphs while it was loading and conductor
(at least the session I was in) locked. I ran
DB.DBA.SPARQL_SELECT_KNOWN_GRAPHS() in isql and it locked my isql
session (apparently) but eventually this came out:
SQL> DB.DBA.SPARQL_SELECT_KNOWN_GRAPHS();
GRAPH_IRI
VARCHAR
_______________________________________________________________________________
http://www.openlinksw.com/schemas/virtrdf#
=======================================
I ran:
select count(*) as ?cc where {?s ?p ?o}
in a sparql session and it came back quick (ie the system is still loading)
I did a status() in another isql and caught what looked like a bunch of
errors in the form:
24498: IER NO OWNER
24499: IER NO OWNER
24500: IER 1
24501: IER NO OWNER
24502: IER NO OWNER
24503: IER NO OWNER
24504: IER 1
24505: IER 1
62: ISR NO OWNER
24506: IER NO OWNER
24507: IER 1
24508: IER NO OWNER
24509: IER NO OWNER
24510: IER NO OWNER
24511: IER 1
24512: IER 1
40: ISR NO OWNER
38: ISR NO OWNER
39: ISR NO OWNER
24513: IER NO OWNER
24514: IER NO OWNER
24515: IER 1
79: ISR NO OWNER
24516: IER NO OWNER
24518: IER NO OWNER
24519: IER NO OWNER
24520: IER 1
74: ISR NO OWNER
24521: IER NO OWNER
24522: IER NO OWNER
24523: IER 1
40: ISR NO OWNER
41: ISR NO OWNER
24524: IER NO OWNER
24525: IER NO OWNER
24526: IER 1
50: ISR NO OWNER
24527: IER NO OWNER
24528: IER 1
41: ISR NO OWNER
24529: IER 1
20: ISR NO OWNER
19: ISR NO OWNER
24530: IER NO OWNER
24531: IER NO OWNER
24532: IER 1
50: ISR NO OWNER
51: ISR NO OWNER
24533: IER 1
47: ISR NO OWNER
Client 1111:1: Account: dba, 384 bytes in, 544 bytes out, 1 stmts.
PID: 32488, OS: unix, Application: unknown, IP#: 127.0.0.1
Transaction status: PENDING, 1 threads.
Locks: 24382: IE, 24343: IS, 24404: IE, 24422: IE, 24483: IE, 24359: IE,
24347: IS, 24410: IE, 24372: IE, 24520: IE, 24351: IE, ....
Client 1111:10: Account: dba, 550 bytes in, 942792 bytes out, 1 stmts.
PID: 32510, OS: unix, Application: unknown, IP#: 127.0.0.1
Transaction status: PENDING, 1 threads.
Locks:
Running Statements:
Time (msec) Text
165 status()
418830 DB.DBA.SPARQL_SELECT_KNOWN_GRAPHS()
Hash indexes
244 Rows. -- 167 msec.
but most got clipped because my putty didn't show history. Once the
load finished, my conductor session had finally flipped to another
screen (I had clicked on a different menu, but when I went back to the
graph list it froze again. I ran:
select distinct ?g where {graph ?g {?s ?p ?o}}
and it timed out like trials #1 and #2. The currently "locked" isql
session just spitted out the loader completions and then spit out two
more lines of the DB.DBA.SPARQL_SELECT_KNOWN_GRAPHS() function:
http://localhost:8890/DAV/
http://www.w3.org/2002/07/owl#
It seems, if you hit the graph listing during the 8 threaded load.
something gums up, but, when you don't, it works fine. I hope this
helps. - Erich
On 09/23/13 7:26 AM, Hugh Williams wrote:
> Hi Erich,
>
> What does the DB.DBA.SPARQL_SELECT_KNOWN_GRAPHS() function return if run as
> this point:
>
> http://docs.openlinksw.com/virtuoso/fn_sparql_select_known_graphs.html
>
> What datasets are you loading when this issue occurs as if they can be
> provided we would like to attempt to recreate in-house ?
>
> 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 23 Sep 2013, at 02:16, Erich Bremer <er...@ebremer.com> wrote:
>
>> The CPU utilization for the virtuoso-t process dropped back down once I
>> returned after a while. As a quick experiment, I ran the SPARQL query:
>> select distinct ?g where {graph ?g {?s ?p ?o}}
>> which drove virtuoso-t to 100% utilization which then timed out returning
>> the CPU back to baseline. Trying to list the graphs in Conductor-->Linked
>> Data-->Graphs which seems to be the same thing (listing all of the named
>> graphs) but "hangs" conductor much longer. Perhaps, there is no timeout on
>> this mechanism on this conductor mechanism. Like before, shutting down
>> virtuoso and then restarting, both the conductor listing of graphs and the
>> above SPARQL query execute in approximately 1 second for the 120,000,000
>> triples and successfully list the 4 named graphs in the system which is the
>> response time that I typically get with Virtuoso, but not until after the
>> restart. - Erich
>>
>> Erich Bremer
>> http://www.ebremer.com
>> http://haylyn.io
>>
>> On 9/22/2013 8:32 AM, Hugh Williams wrote:
>>> Hi Erich,
>>>
>>> Is this condition reproducible as a restart of the server seems to have
>>> resolved it ? As it would be interesting to see what the "status();"
>>> command reports as to the status of the Virtuoso server when in this state
>>> ...
>>>
>>> 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 22 Sep 2013, at 05:32, Erich Bremer <er...@ebremer.com> wrote:
>>>
>>>> Hi all,
>>>>
>>>> I'm working with Virtuoso 7.0.0 open source edition and did a load of
>>>> geonames data using isql which was 123,020,822 triples into a named
>>>> graph "http://www.geonames.org";.
>>>>
>>>> I did a second load of 13,349,346 triples into a different named graph.
>>>> Both loads of turtle RDF completed quickly. When I went to check to see
>>>> if Conductor could see the named graphs, Conductor hung when I clicked
>>>> on Linked Data-->Graphs. I then tried to do:
>>>>
>>>> a) select distinct ?g where {graph ?g {?s ?p ?o}}
>>>>
>>>> from the SPARQL endpoint and it timed out.
>>>>
>>>> however, the following query against http://www.geonames.org would work
>>>> in a second:
>>>>
>>>> b) select count(*) as ?cc where {?s ?p ?o}
>>>>
>>>> I initiated a shutdown using isql which took a few moments and then I
>>>> restarted virtuoso. Now, Conductor quickly shows all named graphs
>>>> including the new ones, and the above SPARQL query (a) runs nearly
>>>> instantly.
>>>>
>>>> - Erich
>>>>
>>>> ------------------------------------------------------------------------------
>>>> LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99!
>>>> 1,500+ hours of tutorials including VisualStudio 2012, Windows 8,
>>>> SharePoint
>>>> 2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack
>>>> includes
>>>> Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/22/13.
>>>> http://pubads.g.doubleclick.net/gampad/clk?id=64545871&iu=/4140/ostg.clktrk
>>>> _______________________________________________
>>>> Virtuoso-users mailing list
>>>> Virtuoso-users@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/virtuoso-users
>> ------------------------------------------------------------------------------
>> LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99!
>> 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint
>> 2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack
>> includes
>> Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13.
>> http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk_______________________________________________
>> Virtuoso-users mailing list
>> Virtuoso-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/virtuoso-users
------------------------------------------------------------------------------
LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99!
1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint
2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack includes
Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13.
http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk
_______________________________________________
Virtuoso-users mailing list
Virtuoso-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/virtuoso-users
