Hi Adam,
the server specs you posted are not so important. What disks did you use?
They should be SSD or 15k RPM SAS to make it faster.
Virtuoso can parse multi thread if you split the files before loading, but hdd speed is still the bottleneck.
Sebastian
Yep!
And if the shared-nothing cluster edition is in use, you can run the bulk loaders in parallel across each of the nodes in the clusters which will reduce the load time too.
We've cluster configurations behind our LOD instance where all of
DBpedia was loaded in 15 minutes flat, and I don't mean via some
massive cluster setup just what we have behind our LOD Cloud cache
instance :)
Kingsley
On June 20, 2019 2:37:16 PM GMT+02:00, Adam Sanchez <a.sanchez75@gmail.com> wrote:For your information a) It took 10.2 days to load the Wikidata RDF dump (wikidata-20190513-all-BETA.ttl, 379G) in Blazegraph 2.1.5. The bigdata.jnl file turned to be 1.3T Server technical features Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 16 On-line CPU(s) list: 0-15 Thread(s) per core: 2 Core(s) per socket: 8 Socket(s): 1 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 79 Model name: Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz Stepping: 1 CPU MHz: 1200.476 CPU max MHz: 3000.0000 CPU min MHz: 1200.0000 BogoMIPS: 4197.65 Virtualization: VT-x L1d cache: 32K L1i cache: 32K L2 cache: 256K L3 cache: 20480K RAM: 128G b) It took 43 hours to load the Wikidata RDF dump (wikidata-20190610-all-BETA.ttl, 383G) in the dev version of Virtuoso 07.20.3230. I had to patch Virtuoso because it was given the following error each time I load the RDF data 09:58:06 PL LOG: File /backup/wikidata-20190610-all-BETA.ttl error 42000 TURTLE RDF loader, line 2984680: RDFGE: RDF box with a geometry RDF type and a non-geometry content The virtuoso.db file turned to be 340G. Server technical features Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 12 On-line CPU(s) list: 0-11 Thread(s) per core: 2 Core(s) per socket: 6 Socket(s): 1 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 63 Model name: Intel(R) Xeon(R) CPU E5-1650 v3 @ 3.50GHz Stepping: 2 CPU MHz: 1199.920 CPU max MHz: 3800.0000 CPU min MHz: 1200.0000 BogoMIPS: 6984.39 Virtualization: VT-x L1d cache: 32K L1i cache: 32K L2 cache: 256K L3 cache: 15360K NUMA node0 CPU(s): 0-11 RAM: 128G Best, Le mar. 4 juin 2019 à 16:37, Vi to <vituzzu.wiki@gmail.com> a écrit :V4 has 8 cores instead of 6. But well, it's a server grade config on purpose! Vito Il giorno mar 4 giu 2019 alle ore 16:32 Guillaume Lederrey <glederrey@wikimedia.org> ha scritto:On Tue, Jun 4, 2019 at 3:14 PM Vi to <vituzzu.wiki@gmail.com> wrote:AFAIR it's a double Xeon E5-2620 v3. With modern CPUs frequency is not so significant.Our latest batch of servers are: Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz (so v4 instead of v3, but the difference is probably minimal).Vito Il giorno mar 4 giu 2019 alle ore 13:00 Adam Sanchez <a.sanchez75@gmail.com> ha scritto:-- Guillaume Lederrey Engineering Manager, Search Platform Wikimedia Foundation UTC+2 / CESTThanks Guillaume! One question more, what is the CPU frequency (GHz)? Le mar. 4 juin 2019 à 12:25, Guillaume Lederrey <glederrey@wikimedia.org> a écrit :On Tue, Jun 4, 2019 at 12:18 PM Adam Sanchez <a.sanchez75@gmail.com> wrote:Hello, Does somebody know the minimal hardware requirements (disk size and RAM) for loading wikidata dump in Blazegraph?The actual hardware requirements will depend on your use case. But for comparison, our production servers are: * 16 cores (hyper threaded, 32 threads) * 128G RAM * 1.5T of SSD storageThe downloaded dump file wikidata-20190513-all-BETA.ttl is 379G. The bigdata.jnl file which stores all the triples data in Blazegraph is 478G but still growing. I had 1T disk but is almost full now.The current size of our jnl file in production is ~670G. Hope that helps! GuillaumeThanks, Adam-- Guillaume Lederrey Engineering Manager, Search Platform Wikimedia Foundation UTC+2 / CEST
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
_______________________________________________ Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Home Page: http://www.openlinksw.com Community Support: https://community.openlinksw.com Weblogs (Blogs): Company Blog: https://medium.com/openlink-software-blog Virtuoso Blog: https://medium.com/virtuoso-blog Data Access Drivers Blog: https://medium.com/openlink-odbc-jdbc-ado-net-data-access-drivers Personal Weblogs (Blogs): Medium Blog: https://medium.com/@kidehen Legacy Blogs: http://www.openlinksw.com/blog/~kidehen/ http://kidehen.blogspot.com Profile Pages: Pinterest: https://www.pinterest.com/kidehen/ Quora: https://www.quora.com/profile/Kingsley-Uyi-Idehen Twitter: https://twitter.com/kidehen Google+: https://plus.google.com/+KingsleyIdehen/about LinkedIn: http://www.linkedin.com/in/kidehen Web Identities (WebID): Personal: http://kingsley.idehen.net/public_home/kidehen/profile.ttl#i : http://id.myopenlink.net/DAV/home/KingsleyUyiIdehen/Public/kingsley.ttl#this