Open Branovitskiy opened 5 years ago
Hi, everybody,
We seem to have scraped the culprit query out of core dump file.
It looks like this:
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX crm2: <http://sp7.ru/ontology/>
SELECT DISTINCT ?gridRelatedUri ?client
WHERE { ?gridRelatedUri <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> ?sparql_type_4532
.?sparql_type_4532 <http://sp7.ru/ontology/subClassOfCache> <http://sp7.ru/ontology/Organization> .
?client <http://sp7.ru/ontology/client> ?gridRelatedUri .
?client <http://sp7.ru/ontology/doer> ?clientDoer .
?client <http://sp7.ru/ontology/in_position> ?clientPosition .
?clientPosition <http://www.w3.org/2000/01/rdf-schema#label> ?clientPositionLabel .
{
SELECT ?clientDoer ?dateBirthDateLabel ?year2 WHERE {
OPTIONAL {
?dateBirth rdf:type crm2:DateAct.
?dateBirth crm2:about ?clientDoer.
?dateBirth crm2:act ?dateBirthAct.
?dateBirthAct rdfs:label ?dateBirthActLabel.
?dateBirth crm2:date ?dateBirthDate.
?dateBirthDate rdfs:label ?dateBirthDateLabel.
?dateBirthDate crm2:sortdate ?dateBirthDateSortdate.
}
FILTER(STR(?dateBirthActLabel) = ‘Дата рождения’)
BIND((now()) - xsd:datetime(?dateBirthDateSortdate) as ?vozrast)
BIND((?vozrast / 31536000) as ?year)
BIND(FLOOR(?year) as ?year2)
}
}
FILTER (!isBLANK(?gridRelatedUri))
FILTER (STR(?clientPositionLabel) = “ГЕНЕРАЛЬНЫЙ ДИРЕКТОР“)
FILTER (?gridRelatedUri = <http://sp7.ru/ontology/res_350417>)
} LIMIT 10
The problem is that it doesn't kill the database by itself.
We have several theories when it happens: 1) In combination with other queries (when there is high load on database) 2) It kills the database only with certain compilation result (or maybe in the process of compilation)
Maybe you could give us some recommendations - what to do next.
Thank you, Stas
@Branovitskiy: When database is under load prior to crash, what does the output of running the status();
command from the isql
command line tool report as to the general status of the database at that point?
@HughWilliams Hugh, hello
Thank you! I will answer as soon as possible. Currently we have problems with kibana logs. :( Working on that.
Hugh, hi.
Sorry, it took us so long.
Here is a link to complete status()
https://drive.google.com/open?id=1UDCticUq6nZSU-0MQ0NZrE85uLCCH7XT
And this is the beginning of status()
OpenLink Virtuoso Server
Version 07.20.3230-pthreads for Linux as of May 27 2019
Started on: 2019-06-24 11:47 GMT+0
Database Status:
File size 1738539008, 25378048 pages, 6940187 free.
20710000 buffers, 11318143 used, 4081 dirty 2 wired down, repl age 0 0 w. io 1 w/crsr.
Disk Usage: 11333796 reads avg 0 msec, 0% r 0% w last 59 s, 1479546 writes flush 210 MB/s,
57591 read ahead, batch = 192. Autocompact 3258 in 3220 out, 1% saved col ac: 1074464 in 0% saved.
Gate: 145360 2nd in reads, 0 gate write waits, 0 in while read 0 busy scrap.
Log = /var/lib/virtuoso-opensource-7/db/trxlog/virtuoso20190625102601.trx, 136859 bytes
196616 pages have been changed since last backup (in checkpoint state)
Current backup timestamp: 0x7D0F-0x1F-0x32
Last backup date: Mon Jun 24 23:05:02 2019
Clients: 220490 connects, max 212 concurrent
RPC: 5526516 calls, -220428 pending, 5 max until now, 0 queued, 477 burst reads (0%), 48 second 14M large, 10947M max
Checkpoint Remap 238431 pages, 0 mapped back. 1517 s atomic time.
DB master 25378048 total 6940187 free 238431 remap 1519 mapped back
temp 26880 total 26875 free
Lock Status: 0 deadlocks of which 0 2r1w, 69 waits,
Currently 3 threads running 0 threads waiting 0 threads in vdb.
Pending:
Hopefully it can give you some hints.
Have a great day! Stas
@HughWilliams
A kind reminder in case our two segfaults (#850 and #858) are connected through the same problem.
Good day, Dear Virtuoso,
Several days ago we started receiving “segmentation fault” errors (
signal 11
) every 5-15 minutes.Our website https://wikilect.ru/
We are using Virtuoso Open Source Edition (Column Store) (multi threaded)
7.2.6-rc1.3230-pthreads
.Rebuild database from TRX (
+crash-dump
and+restore-crash-dump
) had no positive effect.Last 500 records from
virtuoso.log
https://drive.google.com/open?id=1WXQ81Eo2Xya6QR3zMx6QEP_jpeg9TsK6Current config file (
virtuoso.ini
) https://drive.google.com/open?id=13HWYftew4c31KgyTLuPMR4v4iAlc62b0Core dump was generated by
virtuoso-t +foreground
. Program terminated with signalSIGSEGV
, Segmentation fault. Withgdb
we opened core dump file and saw the following:Thank you for any help! Stas