[hub.docker.com] gitea/gitea:1.6.0 webserver not starting in kubernetes.

[Cellebyte]

• Gitea version (or commit ref):
• Git version:
• Operating system:
• Database (use [x]):
  • PostgreSQL
  • MySQL
  • MSSQL
  • SQLite
• Can you reproduce the bug at https://try.gitea.io:
  • Yes (provide example URL)
  • No
  • Not relevant
• Log gist:

Description

Using the hub.docker.com gitea:1.6.0 Image with a custom Dockerfile.

./gitea web

Is not starting up. Building gitea 1.6.0 from source and building a custom docker image of it also not helps. If I go into the container and run the frontend manually via ./gitea web it works as expected. (Busybox Issue?).

FROM gitea/gitea:1.6.0
# Build ontop
RUN apk add --update --no-cache \
    ruby \
    sed \
    ca-certificates \
    asciidoctor && \
    gem install --no-ri --no-rdoc coderay && \
    gem update --no-ri --no-rdoc asciidoctor && \
    gem cleanup && \
    rm -rf /tmp/* /var/cache/apk/*

COPY asciidoc.sh /usr/bin/asciidoc.sh
RUN chown 1000:1000 /usr/bin/asciidoc.sh && \
    chmod 0755 /usr/bin/asciidoc.sh

COPY custom-CA.cer /usr/local/share/ca-certificates/custom-CA.cer
RUN update-ca-certificates

Using the hub.docker.com gitea/latest Docker Image with a custom Dockerfile does not have any of the symptoms provided above but breaks on #5458.

Log File Output

No webserver startup.

Dec  4 10:41:13 syslogd started: BusyBox v1.27.2
Dec  4 10:41:13 sshd[15]: Server listening on :: port 22.
Dec  4 10:41:13 sshd[15]: Server listening on 0.0.0.0 port 22.

[Cellebyte]

#4425 is closed and fixed this issue. A backport to 1.6 would be nice because it breaks on larger installations of gitea in kubernetes. We are using network filesystem storage (GlusterFS). It takes up to 30 minutes to start the pod if it will be scheduled on another host.