# Running Virtuoso

## Tags

* type: documentation
* status: unclear

## Description

We run instances of Virtuoso for our graph databases. Virtuoso is a remarkable software and running some really large databases, including Uniprot

=> https://github.com/openlink/virtuoso-opensource
=> https://www.uniprot.org/sparql/

On Penguin2 virtuoso is running by default as a shepherd servive, see

=> ./shepherd.gmi

```
guix environment --ad-hoc virtuoso-ose -- virtuoso-t -f
```

=> https://git.genenetwork.org/efraim/shepherd-services/src/branch/master/shepherd/init.d/virtuoso.scm penguin2:/home/shepherd/shepherd-services/shepherd/init.d/virtuoso.scm

The database is initialized from 'penguin2:/export/virtuoso/var/lib/virtuoso/db/virtuoso.ini'

### Running a new instance

For testing we want to run a temporary instance. Let's start from the virtuoso.ini file:

```
mkdir -p ~/services/virtuoso
cd services/virtuoso
cp /export/virtuoso/var/lib/virtuoso/db/virtuoso.ini .
```

and edit it to change paths and ports - use non-priviliged ports(!). A full diff is below. Start the server in a screen or tmux (it may ask for creating ./db):

```
penguin2:~/services/virtuoso$ ~/.config/guix/current/bin/guix environment --ad-hoc virtuoso-ose glibc-locales

penguin2:~/services/virtuoso$ /gnu/store/9aqd4jmkafhkdm095hnmxpxzws3ym3wd-virtuoso-ose-7.2.5/bin/virtuoso-t +foreground +configfile virtuoso.ini

03:34:50 HTTP/WebDAV server online at 28890
03:34:50 Server online at 21111 (pid 57078)
```

Now the server should respond to

```
curl localhost:28890/sparql
```

and the admin interface on

```
curl localhost:28890/conductor
```

To use the service from your remote machine use ssh tunnels:

```
ssh -L 28890:127.0.0.1:28890 -f -N myname@penguin2.genenetwork.org
```

and surf to http://localhost:28890/conductor. A good time to change the default password (dba:dba)!

### Uploading data with CURL

To upload RDF I use rapper to validate the data. First delete the existing graph with something like

```
curl -v --digest --user dba:password --verbose --url -G http://localhost:28890/sparql-graph-crud-auth --data-urlencode graph=https://BioHackrXiv.org/graph -X DELETE
```

Next update the graph with

```
curl -v -X PUT --digest -u dba:password -H Content-Type:text/turtle -T test/data/biohackrxiv.ttl -G http://localhost:28890/sparql-graph-crud-auth --data-urlencode graph=https://BioHackrXiv.org/graph
```

Where BioHackrXiv is the name of the graph (in this example). A python version can be found in

=> https://github.com/pubseq/bh20-seq-resource/blob/master/scripts/update_virtuoso/check_for_updates.py

### Virtuoso.ini

What changed in $HOME/services/virtuoso/virtuoso.ini

```
+DatabaseFile                   = $HOME/services/virtuoso/db/virtuoso.db
+ErrorLogFile                   = $HOME/services/virtuoso/db/virtuoso.log
+LockFile                       = $HOME/services/virtuoso/db/virtuoso.lck
+TransactionFile                = $HOME/services/virtuoso/db/virtuoso.trx
+xa_persistent_file             = $HOME/services/virtuoso/db/virtuoso.pxa
+DatabaseFile                   = $HOME/services/virtuoso/db/virtuoso-temp.db
+TransactionFile                = $HOME/services/virtuoso/db/virtuoso-temp.trx
-ServerPort                     = 1111
+ServerPort                     = 21111
+NumberOfBuffers                = 340000
+MaxDirtyBuffers                = 250000
+ServerPort                     = 28890
+ServerRoot                     = $HOME/services/virtuoso/vsp
```