Skip to Content
DocumentationFAQs

Frequently Asked Questions

How can I start a fresh database?

Stop and remove the existing container (including its volumes):

docker stop sensordb
docker rm -v sensordb

A task is too slow, uses too much memory, or the database crashes — what can I do?

Performance depends on both the database and the companion tool (import/export).

  • If both run on the same machine, they compete for CPU and RAM.
  • If they run on different machines, network bandwidth and latency are usually the bottleneck.

Quick guidelines:

  • Start conservative: set --db-max-connections to 5–10.
  • If throughput is low and resources are available:
    • Same machine: gradually increase --db-max-connections (e.g., +5 at a time).
    • Different machines: first check network throughput/latency; increasing connections helps only if the network isn’t saturated.
  • If you see high memory or crashes: lower --db-max-connections.

Tips:

  • Monitor CPU, RAM, disk I/O, and network (throughput, RTT, packet loss).
  • Avoid running other heavy tasks in parallel.
  • On bandwidth-limited links, prefer steady streaming over many aggressive connections.

Goal: use the smallest --db-max-connections that saturates the slowest component (often the network across machines) without exhausting resources.

How can I exclude features from the point cloud association process?

When building models that include both detailed geometries and LOD1 solid geometries, you may want to exclude certain features from the point cloud association process to avoid Z fighting. For example, to exclude all geometries of Building features from the association process, run the following SQL query:

DO
$$
BEGIN
  PERFORM
delete_geometry_data(array_agg(geometry_data.id))
  FROM geometry_data
  LEFT JOIN feature ON feature.id = geometry_data.feature_id
  LEFT JOIN citydb.objectclass ON feature.objectclass_id = objectclass.id
  WHERE objectclass.classname = 'Building';
END$$;

This will enable a distinct association with WallSurface, RoofSurface, and GroundSurface features.

Why do I have network issues with Docker Desktop on Linux?

Docker can be installed on Linux in two main ways:

  • Docker Desktop : Includes a graphical interface, integrated Kubernetes support, and additional developer tools.
  • Docker Engine : A lightweight, command-line-based installation that runs natively on Linux.

While Docker Engine runs directly on the Linux kernel, Docker Desktop runs a virtual machine powered by QEMU to ensure environment consistency across platforms . This extra virtualization layer can lead to network connectivity issues, resource overhead, and reduced performance compared to native Docker Engine.

It is generally recommended to use Docker Engine on Linux, and Docker Desktop on macOS and Windows where virtualization is required.

My client machine is running low on RAM — what can I do?

  • Reduce --db-max-connections — each connection consumes memory on both client and server.
  • Increase swap space — on Linux, follow this tutorial  to add swap. On Windows, adjust the virtual memory size in System Properties → Advanced → Performance Settings.
  • Run the database on a separate machine — offloads memory pressure from the client.
Last updated on
QueriesCitation