Caused by:
PG::StringDataRightTruncation: ERROR: value too long for type character varying(500)
/opt/zammad/app/models/translation.rb:466:in block in to_database' /opt/zammad/app/models/translation.rb:449:in to_database’
/opt/zammad/app/models/translation.rb:322:in block in load_from_file' /opt/zammad/app/models/translation.rb:317:in each’
/opt/zammad/app/models/translation.rb:317:in load_from_file' /opt/zammad/app/models/translation.rb:23:in sync’
/opt/zammad/db/seeds.rb:34:in <top (required)>' /opt/zammad/.rbenv/versions/2.5.8/bin/bundle:23:in load’
/opt/zammad/.rbenv/versions/2.5.8/bin/bundle:23:in `’
Steps to reproduce the behavior:
clean database with rake db: drop and repeat until rake db:seed
Info: Initialisation on a mariadb works without problems.
@MrGeneration if helpful, I can always start the install steps again for testing, but I need advice concerning changes on the ruby stuff as I am more familiar with php and java.
I’m very sorry, I’m afraid I can’t reproduce the issues you describe.
Maybe this has been a temporary problem, to be double check, please fetch the latest stable for Zammad 3.4. Note that this requires ruby 2.6.5 - specific ruby versions can be installed via rvm which might help you.
Please don’t get the following wrong:
If you don’t feel too familar with ruby and source code installations, you may want to switch to the package installation. The package installation does all these manual required steps for you and tries to do as much of your upgrade tasks as possible.
Maybe that’s an alternative that’s the better way for you.
ok, thx for the alternatives, I think I will will try with 3.4. I have a second productive installation with similar environment (proxmox-virtualised), but with mariadb instead of postgres which works fine. But lets see, maybe zammad 3.4 and ruby 3.6 are different.
ruby 2.6.6p146 (2020-03-31 revision 67876) [x86_64-linux] and zammad 3.4 show the same error with. Something is different when running this product in an lxc container with this version of postgres. As this is the only environmental alternative I have, I will stop further tests and look for another ticketing product.
Just for info: we have two zammad instances here, one runs in a proxmox virtualised kvm vm with a remote mariadb as backend - this one is without any problems. The second one needs to be on an lxc container and has a remote postgres 11 database server. Only THIS one is problematic and I have no idea why. And I think its a database issue with postgres, nothing inside the lxc container probably …
All other applications on the proxmox cluster use lxc containers and the database server is a shared service. I have no “hard facts” why I would need kvm for one database and thus “my own database server”
Just for info: the issue could be finally solved. The created database had wrong encoding (SQL_ASCII) instead of UTF8, therefore some of the sql statements had strings beyond the size limit. If you create the database with
create database zammad with encoding 'UTF8' template=template0
all scripts db:migrate and bundle exec rake db:seed can be finished successfully. The encoding info for the target DB should be added into the installation documentation!