Timeout Issue

[anjomshoaa]

Dear qEndpoint team,

We have installed an instance of qEndpoint docker (24G memory) using our existing Wikidata HDT files (not downloaded from your server) which is running here: http://crispy.ai.wu.ac.at/

However the sparql endpoint mostly returns no results with timeout error even for simplest queries. I wonder if it has something to do with the HDT files? Have you made any specific changes in your hosted HDT files?

Thanks!
Amin

[D063520]

Hi,

maybe to start with, what are simplest queries for you? Do they run over the public endpoint? Did you try with our HDTs? We are making modifications to HDT, but currently they should not have a big impact ...

Salut
D063520

[anjomshoaa]

Thanks for the quick reply.

The queries are this simple and they run instantly on Wikidata endpoint:

PREFIX wd: <http://www.wikidata.org/entity/>
SELECT * WHERE {wd:Q84 ?s ?p .}

No, I have not tries your HDT files but before downloading, I wanted to ask and make sure it is the only issue.

Regards, Amin

[ate47]

By checking your logs (http://crispy.ai.wu.ac.at/actuator/logfile) it seems that you have a 5s timeout limit.

10:18:25.710 [http-nio-1234-exec-7] INFO  c.t.q.compiler.SparqlRepository - Running given sparql query: PREFIX bd: <http://www.bigdata.com/rdf#>
PREFIX cc: <http://creativecommons.org/ns#>
PREFIX dct: <http://purl.org/dc/terms/>
PREFIX geo: <http://www.opengis.net/ont/geosparql#>
PREFIX ontolex: <http://www.w3.org/ns/lemon/ontolex#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX p: <http://www.wikidata.org/prop/>
PREFIX pq: <http://www.wikidata.org/prop/qualifier/>
PREFIX pqn: <http://www.wikidata.org/prop/qualifier/value-normalized/>
PREFIX pqv: <http://www.wikidata.org/prop/qualifier/value/>
PREFIX pr: <http://www.wikidata.org/prop/reference/>
PREFIX prn: <http://www.wikidata.org/prop/reference/value-normalized/>
PREFIX prov: <http://www.w3.org/ns/prov#>
PREFIX prv: <http://www.wikidata.org/prop/reference/value/>
PREFIX ps: <http://www.wikidata.org/prop/statement/>
PREFIX psn: <http://www.wikidata.org/prop/statement/value-normalized/>
PREFIX psv: <http://www.wikidata.org/prop/statement/value/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX schema: <http://schema.org/>
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
PREFIX wdata: <http://www.wikidata.org/wiki/Special:EntityData/>
PREFIX wdno: <http://www.wikidata.org/prop/novalue/>
PREFIX wdref: <http://www.wikidata.org/reference/>
PREFIX wds: <http://www.wikidata.org/entity/statement/>
PREFIX wdtn: <http://www.wikidata.org/prop/direct-normalized/>
PREFIX wdv: <http://www.wikidata.org/value/>
PREFIX wikibase: <http://wikiba.se/ontology#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
 PREFIX wd: <http://www.wikidata.org/entity/>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>

SELECT * 
WHERE {
  ?city wdt:P31/wdt:P279* wd:Q515 .
  ?city wdt:P1082 ?pop .
  FILTER (?pop > 1000)
}
ORDER BY DESC(?pop)
10:18:30.779 [http-nio-1234-exec-7] ERROR c.t.q.compiler.SparqlRepository - This exception was caught [org.eclipse.rdf4j.query.QueryEvaluationException: com.the_qa_company.qendpoint.store.exception.EndpointTimeoutException

10:18:25.710
10:18:30.779

Maybe you can try to increase the timeout value. The way we compute the results isn't the same as the one used by Wikidata (Blazegraph) so some queries might not be as fast. For example it seems to take 30s to answer your query.

[D063520]

30 seconds for a simple triple pattern with fixed subject is not normal .....

this is a bit a different setup, but also with an HDT and qendpoint

[ate47]

True, I didn't notice the 1k results. Can we know more about the hardware @anjomshoaa? (Disk/CPU)

Also can you give us the header of the HDT?

# PowerShell (Windows)
gc -TotalCount 19 your-hdt.hdt
# bash (Linux/MacOs)
head -n 19 your-hdt.hdt

[anjomshoaa]

Thanks, the city population query was adopted from one of qEndpoint papers:

Willerval, A., Bonifati, A., & Diefenbach, D. (2023, April). qEndpoint: A Wikidata SPARQL endpoint on commodity hardware. In Companion Proceedings of the ACM Web Conference 2023 (pp. 119-122).

Also regarding the wd:Q84 query: if you insist querying after 4-5 times the results appear and then it will be fast for repeated queries. However, if you look for another entity it is slow again :-(

[anjomshoaa]

True, I didn't notice the 1k results. Can we know more about the hardware @anjomshoaa? (Disk/CPU)

Also can you give us the header of the HDT?

# PowerShell (Windows)
gc -TotalCount 19 your-hdt.hdt
# bash (Linux/MacOs)
head -n 19 your-hdt.hdt

Sure, here is the HDT header:

$HDT<http://purl.org/HDT/hdt#HDTv1>v5$HDTntripleslength=1841;?_:statistics <http://purl.org/HDT/hdt#originalSize> "2666797479600" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://purl.org/HDT/hdt#Dataset> .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://rdfs.org/ns/void#Dataset> .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://rdfs.org/ns/void#triples> "18090254954" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://rdfs.org/ns/void#properties> "51024" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://rdfs.org/ns/void#distinctSubjects> "1931121124" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://rdfs.org/ns/void#distinctObjects> "3320482641" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://purl.org/HDT/hdt#formatInformation> "_:format" .
_:format <http://purl.org/HDT/hdt#dictionary> "_:dictionary" .
_:format <http://purl.org/HDT/hdt#triples> "_:triples" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://purl.org/HDT/hdt#statisticalInformation> "_:statistics" .
<file://[/backup_nfs/hdt-pipeline/download/latest-all.nt]> <http://purl.org/HDT/hdt#publicationInformation> "_:publicationInformation" .
_:publicationInformation <http://purl.org/dc/terms/issued> "2023-05-16T18:45Z" .
_:dictionary <http://purl.org/dc/terms/format> <http://purl.org/HDT/hdt#dictionaryFour> .
_:dictionary <http://purl.org/HDT/hdt#dictionarynumSharedSubjectObject> "1738683562" .
_:triples <http://purl.org/dc/terms/format> <http://purl.org/HDT/hdt#triplesBitmap> .
_:triples <http://purl.org/HDT/hdt#triplesnumTriples> "18090254954" .
_:triples <http://purl.org/HDT/hdt#triplesOrder> "SPO" .
_:statistics <http://purl.org/HDT/hdt#hdtSize> "193188300658" .

also CPU information (on the host where docker runs) is as follows:

Architecture:                       x86_64
CPU op-mode(s):                     32-bit, 64-bit
Byte Order:                         Little Endian
Address sizes:                      48 bits physical, 48 bits virtual
CPU(s):                             96
On-line CPU(s) list:                0-4,10,15,20-22,34,37,38,40,69,90
Off-line CPU(s) list:               5-9,11-14,16-19,23-33,35,36,39,41-68,70-89,91-95
Thread(s) per core:                 0
Core(s) per socket:                 24
Socket(s):                          2
NUMA node(s):                       2
Vendor ID:                          AuthenticAMD
CPU family:                         23
Model:                              49
Model name:                         AMD EPYC 7352 24-Core Processor
Stepping:                           0
Frequency boost:                    enabled
CPU MHz:                            2300.000
CPU max MHz:                        2300.0000
CPU min MHz:                        1500.0000
BogoMIPS:                           4591.83
Virtualization:                     AMD-V
L1d cache:                          768 KiB
L1i cache:                          768 KiB
L2 cache:                           12 MiB
L3 cache:                           128 MiB

Thanks!

[anjomshoaa]

Also I wanted to add that HDT files are on a different server and mounted on the docker host. The mounted folder is used as volume by qEndpoint docker.

[D063520]

Also I wanted to add that HDT files are on a different server and mounted on the docker host. The mounted folder is used as volume by qEndpoint docker.

This is not a good idea for a database, basically the bottleneck is the speed of transfer between the 2 servers. Do you have another option for this?

[anjomshoaa]

Thanks for the feedback. I will move the data to the hosting server then. Hopefully this will solve the problem :-)

[anjomshoaa]

After moving the HDT files to the same server as docker the issue with timeout is solved. Thank you for your support :-)