Environment
- Image:
wiktorn/overpass-api (latest)
- Mode:
init
- Extract:
france-latest.osm.pbf from openstreetmap.fr
OVERPASS_META: yes
OVERPASS_RULES_LOAD: 100
OVERPASS_SPACE: 25769803776 (24 GB)
OVERPASS_UPDATE_SLEEP: 86400
- Server RAM: 32 GB
Problem
The areas_rules process starts and processes elements correctly, but always crashes and restarts around 105,000–128,000 elements. This happens indefinitely — make-area never completes a full cycle.
Observed logs
After 0h5m47s: in "make-area", part 0, on line 29. Stack: 0 of 0 105350 of 0
After 0h6m30s: in "make-area", part 0, on line 29. Stack: 0 of 0 123427 of 0
After 0h0m15s: in "make-area", part 0, on line 29. Stack: 0 of 0 1461 of 0 ← restart
The counter resets to ~1400 each time, consistently after reaching 105k–128k elements.
Troubleshooting done
- No OOM kill visible in
dmesg
- No error in
journalctl
- No
FATAL/exited lines in supervisord logs
OVERPASS_HEALTHCHECK set to exit 0 — rules out healthcheck interference
OVERPASS_UPDATE_SLEEP set to 86400 — rules out diff updates interrupting
OVERPASS_SPACE increased from 12 GB to 24 GB — no change
area(3601403916) returns empty — areas never written to DB
- Same behavior reproduced with Switzerland extract
Expected behavior
areas_rules should complete a full cycle and write areas to the database.
Question
Is there a known relation or element in these extracts that causes make-area to crash?
Environment
wiktorn/overpass-api(latest)initfrance-latest.osm.pbffrom openstreetmap.frOVERPASS_META: yesOVERPASS_RULES_LOAD: 100OVERPASS_SPACE: 25769803776 (24 GB)OVERPASS_UPDATE_SLEEP: 86400Problem
The
areas_rulesprocess starts and processes elements correctly, but always crashes and restarts around 105,000–128,000 elements. This happens indefinitely —make-areanever completes a full cycle.Observed logs
The counter resets to ~1400 each time, consistently after reaching 105k–128k elements.
Troubleshooting done
dmesgjournalctlFATAL/exitedlines in supervisord logsOVERPASS_HEALTHCHECKset toexit 0— rules out healthcheck interferenceOVERPASS_UPDATE_SLEEPset to 86400 — rules out diff updates interruptingOVERPASS_SPACEincreased from 12 GB to 24 GB — no changearea(3601403916)returns empty — areas never written to DBExpected behavior
areas_rulesshould complete a full cycle and write areas to the database.Question
Is there a known relation or element in these extracts that causes
make-areato crash?