Starting with an empty database

[seebi]

We're currently starting qlever in our orchestration by overwriting the entrypoint with the following script where the ${...} variables are injected with helm or docker compose.

Can you suggest a better way for this usecase?

For example, the request-body-limit config option is not available with the start command, so we cannot set it on startup.

#!/usr/bin/env bash
set -Eeuo pipefail

# If the metadata file does NOT exist, build the index.
if [[ ! -f "cmem.meta-data.json" ]]; then
  # Send an empty line to IndexBuilderMain, pipe output to a log file.
  echo '' | IndexBuilderMain \
    --file-format nt \
    --kg-input-file - \
    --index-basename cmem \
    | tee cmem.index-log.txt
fi

# start QLever in the foreground on the provided port/token.
exec qlever start \
--kill-existing-with-same-port \
--port "${QLEVER_PORT}" \
--run-in-foreground \
--persist-updates \
--system native \
--name cmem \
--description "cmem base index" \
--access-token "${QLEVER_ACCESS_TOKEN}"

[hannahbast]

@seebi Starting from qlever version 0.5.32 you can now set runtime parameters by appending them to the qlever start command, as described here: qlever-dev/qlever-control#210

[seebi]

@hannahbast thanks for this. When do you plan to update the docker image so that this is available?

At the moment, 0.5.29 is in the image

∴ docker run --pull always -i -t --rm --entrypoint bash adfreiburg/qlever:latest -c "qlever --version"
latest: Pulling from adfreiburg/qlever
Digest: sha256:60c62c4f77cea7d49110de797df74a060fb6c760cd92c5ccaeac8fed40aae924
Status: Image is up to date for adfreiburg/qlever:latest
qlever 0.5.29

[hannahbast]

@seebi You always find the latest version here https://pypi.org/project/qlever/

In your setup, are you really using the qlever from within the Docker container?

[seebi]

Yes, I am sure. The given entry point script above is mounted inside the container atm. and executed inside of the container.

[hannahbast]

@seebi The typical way how folks use QLever is to do pip install qlever (which will always give you the latest version or you can easily update it to the latest version) and then when you call qlever (with --system = docker, which is also the default) it will use the Docker image you specify. So in that scenario, the qlever command-line tool lives outside of the Docker container and calls docker run ....

My question: Is it an option for you to also use it that way or do you have to use it from within the container?

[seebi]

@hannahbast the container are started either with docker compose or helm (and other kubernetes API clients). A custom client can not be used there.