New fear unlocked: Databases

[nomadscientist]

So I've just heard that tools can require databases, which are not automatically installed with the tool?

i.e. we can have a lovely tool panel for subdomains centralised here, and when a server installs the tools and says 'Look, we now have a Single-cell or Proteomics Lab', the tools may exist but fail because their databases must be also manually installed?

Is this how this works? This is clearly a problem for us...

[paulzierep]

It's one of the main issues with synchronizing Galaxy servers, it's worked on here: https://github.com/galaxyproject/idc
But I think there is not much we can do here, yet. Maybe checking which tools need reference data. But making sure its available on all servers is still a difficult issue.

[nomadscientist]

Can checking which tools need references be automated, for example? So we know which tools to check?

[bebatut]

We could check if the tool suite has a data table in the code source