Closed alpapan closed 1 month ago
NB: galaxy starts fine. i don't know if manage_toolkit db is needed.
providing my experience for manage_db as FYI or for educating me i did something wrong
Your postgres is 12 or less (which I believe it still officially supported by galaxy?) but the migration makes use of a postgres 13+ function, you'll need to CREATE EXTENSION pgcrypto;
to make it available and then re-run the migration.
cc @jdavcs who was talking about this in the admin matrix channel.
Reproduced. Working on a fix.
@alpapan Correction: not reproduced (the similar error I encountered was unrelated).
It seems you are looking at two different databases: the error raised by the manage_toolshed_db.sh
script is a sqlite error (sqlite3.OperationalError
), whereas the database you're checking is postgresql (\d api_keys
). The sqlite error you're seeing will be raised when you run the upgrade/downgrade command on a database that has not been initialized (it will contain only one table: alembic_version). Please check your configuration and let us know if that's indeed the case.
Hello
hexylena is right on the manage_db.sh
. my workaround did the same thing. I can update postgres if needed
I have a community.sqlite
and a results.sqlite
(these are new right?) i don't have a universe.sqlite though
Is it needed? How do i create them? It didn't happen during upgrade
If i don't need to run manage_toolshed_db.sh
then i wont!
# By default, Galaxy uses a SQLite database at
# '<data_dir>/universe.sqlite'. You may use a SQLAlchemy connection
# string to specify an external database instead.
# Sample default
# 'sqlite:///<data_dir>/universe.sqlite?isolation_level=IMMEDIATE'
# You may specify additional options that will be passed to the
# SQLAlchemy database engine by using the prefix
# "database_engine_option_". For some of these options, default values
# are provided (e.g. see database_engine_option_pool_size, etc.).
# The same applies to `install_database_connection`, for which you
# should use the "install_database_engine_option_" prefix.
# For more options, please check SQLAlchemy's documentation at
# https://docs.sqlalchemy.org/en/14/core/engines.html?highlight=create_engine#sqlalchemy.create_engine
#database_connection: null
You'd only need manage_toolshed_db.sh
if you're running your own tool shed.
community.sqlite
is the default sqlite database for tool shed. universe.sqlite
is the default for galaxy. They are set in your config files: galaxy.yml
for galaxy and tool_shed.yml
for tool shed. If a config file is missing, the sample config will be used (e.g. tool_shed.yml.sample
). The config property in both cases is database_connection
. results.sqlite
is the Celery results backend galaxy uses by default. In other words, you only need them if that's what you specified in your configuration files.
Closing this. Please let us know if you run into any further issues.
Galaxy Version and/or server at which you observed the bug Galaxy Version: 24.1 Commit: 2a7954c51bcc2fb6a33879fe157735fb68198f83 upgrade to 2a7954c51b
Browser and Operating System Ubuntu 20.04.6 LTS psql 12
Describe the bug
I noticed that /admin/users (roles and groups) hangs with my current commit even though the relevant API call returns immediately (and no errors in log even with database logging turned on), as per issue https://github.com/galaxyproject/galaxy/issues/18878#issue-2544336463
So I did a git pull on my release_24.1 branch to check if that was an update issue
but the upgrade failed
Issue 1 and solution i did
solution that worked
issue 2, i have not fixed this
The api_keys exists in the database and it already has deleted
FWIW my git pull log
i tried run.sh after git pull but it failed due to database not migrated.