To investigate this further, it would be useful to get a repeatable test case, preferably where the query is failing against a remote Mulgara triplestore.
If anyone is running into this issue with a remote Mulgara instance, a dump of triples (see https://wiki.duraspace.org/display/FCR30/Resource+Index+Search#ResourceIndex Search-SyntaxforRequestingTriples - format N-Triples, with a query pattern of * * *) that could be used to reproduce the problem would be great. Also details of memory configuration both for the Tomcat JVM and the Mulgara JVM, and your OS (particularly if it is 32- or 64-bit). Regards Steve > -----Original Message----- > From: James, Eric [mailto:eric.ja...@yale.edu] > Sent: 05 October 2010 16:59 > To: Durbin, Michael R; fedora-commons-users@lists.sourceforge.net > Subject: Re: [fcrepo-user] OAI provider, or resource index problem > > > Mike, > > We encountered something similar where this query (or our > equivalent of it): > > select $item $itemID $date $state > from <#ri> > where $item > <http://www.openarchives.org/OAI/2.0/itemID> $itemID > and $item > <info:fedora/fedora-system:def/model#state> $state > and $item <info:fedora/fedora-system:def/model#hasModel> $model > and $model <info:fedora/fedora-system:def/model#hasService> $SDef > and $SDef <info:fedora/fedora-system:def/model#definesMethod> 'getDC' > and $SDef <http://mulgara.org/mulgara#is> > <info:fedora/bdef:iudlDescMetadata> > and $item > <info:fedora/fedora-system:def/view#lastModifiedDate> $date > and $date <http://mulgara.org/mulgara#after> > '2010-09-29T19:57:05.983Z'^^<http://www.w3.org/2001/XMLSchema# > dateTime> in <#xsd> > and $date <http://mulgara.org/mulgara#before> > '2010-10-03T22:41:44.034Z'^^<http://www.w3.org/2001/XMLSchema# > dateTime> in <#xsd> > order by $itemID asc > > results in an outOfMemory error in the fedora/risearch trying > to execute this query. We tried separate tomcat instances or > fedora and oaiprovider - no avail. We increased jvm memory - > no avail. Curiously if you remove the "before" and "after" > triples from the where clause, (getting everything, > regardless of when), the query succeeds. > > Anyway, that is as far as we got, please post if you ever get > this to work. > > -Eric > ________________________________________ > From: Durbin, Michael R [midur...@indiana.edu] > Sent: Tuesday, October 05, 2010 11:38 AM > To: fedora-commons-users@lists.sourceforge.net > Subject: [fcrepo-user] OAI provider, or resource index problem > > Fedora users, > > In trying to expose several collections using the oai > provider for fedora (proai) we encounter intermittent > problems that result in the bogging down of our running > fedora repository that persists until the application is > restarted. This happens to the fedora repository even when > the oaiprovider is running in a different tomcat instance on > a different machine. > > What happens is that occasionally the normal resource > index queries that are used to populate the oai database and > cache never return, and appear to result in a thread left > processing forever (the CPU usage remains high, independent > of other operations). At first I thought the problem was > simply a matter of not waiting long enough for the response, > but after increasing the driver.fedora.querySocketTimeout to > 3600 (an hour), query still fails to return on rare occasions. > > Attached is the log file for the oai provider. There was > nothing but unrelated access information in the fedora log. > > I can understand the query taking a long time, but is seems > clear to me that sometimes (possibly times when other > processes are also hitting the resource index) an error > occurs that results in an infinite processing loop. > > We're currently running fedora 3.2. We're using the > localMulgaraTriplestore for the resource index. We're going > to update to fedora 3.4 soon, but given that we can't > reliably reproduce this error (it's happened once in a couple > weeks) it'll be hard to be confident that it's fixed even if > it doesn't occur. > > Have other people experienced this problem? What sort of > log4j configuration would capture more relevant information > about what the resource index is doing? > > Thanks for any information you can provide! > > -Mike Durbin > -------------------------------------------------------------- > ---------------- > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time > creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > Fedora-commons-users mailing list > Fedora-commons-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/fedora-commons-users > ------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ Fedora-commons-users mailing list Fedora-commons-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fedora-commons-users
