Seafile

Seafile 7.0.x to 7.1.x upgrade

Jul 17, 2020 • 2 min read

Install python and other deps (python 3.6 is needed)

Make sure the base install is up to date
apt-get install python3 python3-setuptools python3-pip -y
pip3 install --timeout=3600 Pillow pylibmc captcha jinja2 sqlalchemy django-pylibmc django-simple-captcha python3-ldap

Start upgrade process

wget 7.1.x package
- https://s3.eu-central-1.amazonaws.com/download.seadrive.org/seafile-server_7.1.4_x86-64.tar.gz
stop seafile server
- service seafile-server stop
./upgrade_7.0_7.1.sh
This error is ok and expected on communty edition
- [WARNING] Failed to execute sql: (1091, "Can't DROP 'drafts_draft_origin_file_uuid_7c003c98_uniq'; check that column/key exists")
wipe seahub cache
- rm -rf /tmp/seahub_cache
verify all seafile directories are owned by seafile user
- if not in use already create a user for seafile
- useradd -r -M -s /usr/sbin/nologin

Debugging

Seahub troubleshooting

service seafile-server start failed
- systemctl status seafile-server
- journalctl -xe
- This did not provide as much useful information as running seahub with daemon=False
- Make sure everything stopped running
  - service seafile-server stop
  - ./seafile.sh stop
  - ./seahub.sh stop
Change conf/gunicorn.conf
- daemon = False
- This allows you to manually start seahub.sh (must be run as the user running seafile)
assume seafile user and start seafile/seahub as user
- su -s /bin/bash $user
- ./seafile.sh start
- ./seahub.sh start
Once that runs, you are able to see the error in std out and potential missing modules
- once those modules are fixed and you are able to run seahub manually. Verify all seafile processes are stopped before using service to start the process
  - ./seahub.sh stop
  - ./seafile.sh stop
  - ps aux | grep sea*
  - service seafile-server start

Sources