Closed stonemaster closed 6 years ago
Unfortunately I dind't find a cause why this is happening. Couldn't reproduce it in a clean container. What I did is to start the container with bash
as command line and executed the upgrade scripts in /seafile/seafile-server-latest/upgrade/
manually, while removing the offending lines in /seafile/seafile-server-latest/upgrade/sql/sqlite
until everything passed. Then I changed the /data/seafile_version
file's contents to 6.2
to prevent upgrade scripts with next start. Don't forget to run the following commands after the container starts with bash
as entry point:
rm -rfv /seafile/seahub-data /seafile/seafile-data /seafile/seahub.db \
/seafile/ccnet
ln -s /data/seahub-data /seafile
ln -s /data/seahub.db /seafile
ln -s /data/seafile-data /seafile
ln -s /data/ccnet /seafile/ccnet
This sets up the links to the host mount. Sorry about that. The current Travis build will tests upgrades in future.
Upgrade from a 6.0 container fails:
Looks like an upstream bug. Current workaround is to use the
stonemaster/docker-seafile-server:6.0
container until this is resolved or re-upload data.