Anyone else having issues with Salmon Data Manager never installing?

I run my own local galaxy on my deskstop and I am able to install any tool but for some reason the salmon and kallisto data managers are stuck in the cloning phase for over 12 hours. Is this typical? When I installed featureCounts which comes with its own built in reference genome it was much faster. Any advice would be appreciated!

Welcome, @Jack_Goodman

Do you mean that just installing those two Data Manager tools is taking a long time? Example repository

How long it takes to install tools depends on a few things, including: your connection speed, how busy the ToolShed is, and how many other tools you may be trying to install at the same time.

The Main ToolShed appears to be Ok according to the status page → https://status.galaxyproject.org/

Please clarify and we can follow up more. :hammer_and_wrench:

1 Like

Hi Jenna, I can successfully install any tool from the toolshed including the salmon and kallisto suite of tools but it is their data managers that are unsuccessful. So I don’t think it is my internet connection, and any other tools I have attempted to install at the same time complete.

After clicking install for the salmon or kallisto data manager, I can see the status change to cloning. But despite running for over a day, neither data manger has made it past the cloning stage. I have tried shutting down my galaxy and restarting, stopping the install, resetting the tool and trying to reinstall. But none of that has worked. So I’m at a bit of a loss of what to do at this point.

Hi @Jack_Goodman

Thanks for clarifying. I’m asking our admins for feedback. More soon!

1 Like

Hi Jenna, in playing around with it more I’ve realized its is actually any data manager. I tried using refgenie’s, and that was also unsuccessful. The tools still show up in local data and let me run them but I get this error:

/home/jack/galaxy/database/jobs_directory/002/2255/tool_script.sh: line 10: refgenie: command not found

Hi @Jack_Goodman

Thanks for following up, and posting more details!

I asked internally about this, and others were able to get these loaded, and quickly, without a problem. They are sourcing from the Main ToolShed.

They suggested reviewing how these were being loaded, with guesses about some problem with conda environment.

So – my next questions are: 1) are you using dependency resolvers to install the tools, and 2) have you set up anything different from the default order? What happens if you adjust that order? I’m referring to this → Dependency Resolvers in Galaxy — Galaxy Project 24.1.0.dev0 documentation

And what you posted back is another clue! This can be due to a Python version problem, which also points back to a conda environment mixup.

Thanks Jenna,

I have not altered the dependency resolvers section of the config file so it should be set to default. This is starting to get out of my depth and I don’t have any meaningful data on my local Galaxy. I am happy to delete it and re-install, everything should work with a new reinstall right? I’m running Ubuntu 2204.3.49.0 on Windows and my python version is 3.12. Should I just use this tutorial for installation of Galaxy?
Get Galaxy - Galaxy Community Hub (galaxyproject.org)

Hi @Jack_Goodman

Reinstalling … hard to know if that will be enough.

The GetGalaxy guide is highly simplified … the tutorials are a much better resource. Private Galaxy Servers

Make sure that you are starting up Galaxy the very first time in a Conda venv, and not a Python venv. One of the symptoms of starting up in Python is tools not being found, or not installing. Using Conda instead solves that.

Then, if this happens again, checking the logs and processes running is where you can learn more about why that isn’t resolving. This is more help from our admin Nate:

The tool repo itself could install but then conda spins forever trying to install the dependencies for the tool. They should be able to tell by looking for conda create processes with ps (it can also be in the log, but that gets buried)

We can follow up more, so let us know what you decide to do and what happens! :slight_smile: :hammer_and_wrench:

I think that’s the issue. It’s starting up with a python venv not conda. How would I make the swap?

1 Like

Hi @Jack_Goodman

I don’t think there is documented way to make a switch after. Why? There are too many items that are set up automatically when Galaxy is first launched. Those are slightly different each release, and between any two particular releases, plus any post-launch modifications.

Now, others can correct me, and you could ask at the Admin matrix chat to confirm. Maybe there is something “more advanced” that could be done. Let us know how it goes! It is a pain to start over but every time we’ve pursued this, it always rolls back to that.