Routino database creation fails: Fatal Error (files.c:610): Cannot write to file .... relationsx.route......tmp' [Bad address].

[wolfgangr]

Describe the bug

QMapShack fails to create a routino database from current europe-latest.osm.pbf

Seems to be a routino issue.

Both in the QMapShack window and directly from the command line I get

~/tracks/qMapShack-Karten/routino-DB$ /usr/bin/planetsplitter --dir=/home/username/tracks/qMapShack-Karten/routino-DB --prefix=Europe_26 --tagging=/usr/share/routino/tagging.xml --keep --sort-ram-size=80000  --sort-threads=18  /home/username/tracks/qMapShack-Karten/routino-DB-26/europe-latest.osm.pbf

Parse OSM Data [/home/username/tracks/qMapShack-Karten/routino-DB-26/europe-latest.osm.pbf]
==============

Reading: Bytes=34150026316 Nodes=3784109501 Ways=462863577 Relations=413000Routino Fatal Error (files.c:610): Cannot write to file '/home/username/tracks/qMapShack-Karten/routino-DB/relationsx.route.0x560e3c56a2f0.tmp' [Bad address].

.... la ....
37261090 Jul 12 17:48 relationsx.route.0x560e3c56a2f0.tmp

edit: above is a manual call from the cmd line.
For proof, I repated the call from QMS again.
It produces the same error, at the same counters, except a different random in the tmp file name.
Params for planetsplitter differ slightly:

/usr/bin/planetsplitter --dir=/home/username/tracks/qMapShack-Karten/routino-DB --prefix=EUR26 --tagging=/usr/share/routino/tagging.xml --parse-only /home/username/tracks/qMapShack-Karten/routino-DB-26/europe-latest.osm.pbf

What have you done to circle down the problem?

• tried to repeat the database generation from the command line
• checked for current Routino version 3.4.3
• web and forum searches,
• found https://www.routino.org/documentation/limits.html
  could not find precise the same issue reported there nor anywhere else
• I'm still working with a routino database created long ago back in 2017

To Reproduce

1. Download https://download.geofabrik.de/europe-latest.osm.pbf
   as of 2026-07-12
2. in QMS, open Tool/Create Routino Database
3. select source and target path, prefix as effective in the error log above
4. click start and watch the process working
5. after about an hour, find the error message reported

Expected behavior

planetsplitter finishing and producing workable database

Screenshots

no screenshots contributing relevant information

Attachments

don't dare to attach 33 G file:
33G Jul 12 02:16 europe-latest.osm.pbf

Tracebacks

n.a.

Desktop

• OS: debian trixie 13.4
• QMapShack Version: 1.17.1
• Window Manager: lxde

Additional context

QMapShack and routino are working with the old 2017 database over serveral version upgrades both on linux and windows.

I'm considering a custom build of routino with increased index limits.
How is the search built into QMS? By command line? By library? Can I have concurrent versions of routino (debian and custom build) on the machine, e.g. in ~/.local/bin ?

Whats the proper way to report an issue with routino?

[kiozen]

Imho this is nothing we can fix in QMapShack. This is a problem with the build tools of Routino. My guess: the task is getting too large. But you have to discuss that with the author of Routino.

[JFritzle]

In December 2025, I could succesfully build a Routino database from whole Europe using
europe-latest.osm.pbf, suitable for Osmium, Osmosis, imposm, osm2pgsql, mkgmap, and others.
This file was last modified 8 hours ago and contains all OSM data up to 2025-12-03T21:21:00Z.
File size: 31.2 GB; MD5 sum: a0ac69f95ce000139dadc20d32ab2567.
On an old 4 cores PC (-> 4 threads!) with 24 GB RAM it took about 50 hours and resulted in a database of 11.6 GB.

My build script planetsplitter.zip uses 2 steps:

• step 1: parse-only
• step 2: process-only

If desired, I could upload database into cloud and provide a download link valid for a limited short time.

[wolfgangr]

discuss that with the author of Routino

I know. Sent a mail to AMB already. Pointed him to this issue, so I hope to welcome him here, soon.

My guess: the task is getting too large.

That was my first guess, too.
But then I'd expect the failure at some point in the process where large figures are hit.
However, the huge amounts of nodes and ways is already processed at this stage.
And the error message does not indicate an overrun of sth. like index space.

Fatal Error (files.c:610)

Gleaning at routino source, this looks like some low level buffer writing code.
Maybe skd of buffer overflow?
May be triggered by faulty/unusual input data?
May be triggered by some opaque side effect?
@ amb ?

---

This all begs the question:
Is routino still the tool of choice for Enduro Travel routing?

Long cross country (and maybe even cross continent) tours including small, ternary and below, roads?
Ending up with huge data sets, even if used only sparsely, steadyly growing with proceeding detailed survey of OSM.
Appears to me, that the concept of a single prebuilt index hits a ceiling?

Thus I took this as a kick to give BRouter a second try with local installation and on-the-fly routing.
After initial setup quirks, I'm impressed.

My old tile selection (whole Europe, North Africa and half of Asia, based on silk road dreams) was updated within a couple of minutes.
Small tracks in my local vicinity are mapped quite well.
Recent road changes are up to date.
Long routes ( ~ 1500 km, segments ~ 300 km ) become clumsy (slower than routino) but still manageable.

@kiozen, how would you asses "routino vs BRouter" from a skd of strategic perspective?

[kkarsten62]

I code a small bash script build_routino-DB.sh to build one routino database for some selected European countries. But I am always on latest QMS development commit, currently based on version V_1.20.3.
Copy it to the routinoDbDir folder. Maybe it will help and you can adapt it for your own needs:

#!/bin/bash

url='https://download.geofabrik.de/europe/'
fileExtension='-latest.osm.pbf'
planetsplitter='/usr/bin/planetsplitter'
routinoDbDir='/home/karl/GPS/routino-DB/'
taggingDir='/usr/share/routino/'
prefix='EU'

# Structure is:
# 'filename w/o extension'
datas=(
'austria,'
'belgium,'
'france,'
'germany,'
'italy,'
'liechtenstein,'
'luxembourg,'
'netherlands,'
'switzerland,'
)

no="N"
read -p $'Is the folder empty - except the build script, y/N ? ' no
if [[ "$no" != "y" && "$no" != "Y" ]]; then
	exit
fi

for data in "${datas[@]}"
do
	pbfFile=$(echo "${data}" | awk -F',' '{ print $1 }')
	echo -e "\n\n --- Downloading and parsing file for region: ${pbfFile} ---"
	wget ${url}${pbfFile}${fileExtension}
	${planetsplitter} --dir=${routinoDbDir} --prefix=${prefix} --tagging=${taggingDir}'tagging.xml' --parse-only --append --sort-threads=$(nproc) ${pbfFile}${fileExtension}
done

echo -e "\n\n --- Process and build DB ---"
${planetsplitter} --dir=${routinoDbDir} --prefix=${prefix} --tagging=${taggingDir}'tagging.xml' --process-only --sort-threads=$(nproc)
echo -e "\n --- Finished ---"

[wolfgangr]

@JFritzle

I could upload database into cloud and provide a download link valid for a limited short time.

Great, that would postpone the need for a solution for me for another half of a decade :-)
If it's little effort for you, I'd gladly accept the gift.

Otherwise, I'd try @kkarsten62's script and build subregion databases.

[JFritzle]

@wolfgangr

This download link for Routino database Europe is valid for 1 week.