History completely disappeared from local galaxy after relaunching the instance

Dear All,

I have a local Galaxy instance. I had a history with about 500Gb data, all analyses were finished. I had to restart the computer and relaunch Galaxy. Unfortunately, the history completely disappeared, and the “using 500 Gb” became “using 0”. This has not happened before.
This time during the relaunch of Galaxy, however I got a message that the pip needed to be updated, which I did.
To my knowledge files corresponding to that history were still in the database/object folder, but were somehow not detected by the system.
Any idea what might have happened? Was it because pip

Thanks in advance!

Jean-Michel Fustin


Your history is stored in the database.
Are you using the default history created when the user first logs in?

Check that nothing happened or was reconfigured relating to the database.

1 Like

Agree with @innovate-invent

One thing to note is that if you were not working in a logged-in account, then the data might actually be lost, or at least lost to the Galaxy application. Only accounts that are logged in save history data (and, you can only have one history at a time).

If you were logged in, maybe you are logging in to a different account than before?

The data may still be on the disc – but detached from Galaxy. It depends on how you have set up the local configuration. It sounds like you still have it (did not run any admin data purging processes).

If you need more help – please let us know and we can bring in some more admin advice. There may be a workaround.

Hello @jennaj and @innovate-invent
Thank you for your input.
For now, as I am learning, I have activated the single_user option, and of course the admin_users with the same email address.
Will histories be saved in this case?

1 Like

Were you logged into that account when doing the work?

If yes, please share back:

  1. The version(s) of Galaxy you were running originally then upgraded to (if upgraded)
  2. Where/when it was sourced for the original and upgrade
  3. How the upgrade was done
  4. Which “single_user” option(s) you modified and when (original or after upgrading?). This is in a few places. https://github.com/galaxyproject/galaxy/search?q=single_user&unscoped_q=single_user
  5. Check to see if you have two accounts – possibly there is a case-sensitive double account issue going on.

Example: These could be two distinct accounts, it depends on other config options.


GUI function that is not case-sensitive, and useful. Maybe query your email address, or part of it, and make sure there are not two (or more) accounts with case as the only difference. This includes deleted/purged accounts – which may or may not show up in the GUI but could be in the database (you can query it that way as well). If you turned on the GDPR mode then purged accounts have the email anonymized. Custom scripts or queries could be problematic if “case” is not considered.

Admin > Users > Search

Let’s troubleshoot more from there :slight_smile:

1 Like

Hi @jennaj
Thanks for helping.
Well, it is my understanding that since the “single_user” option is active I am always logged in as soon as I launch Galaxy. Indeed “logged in as xxxx” is always displayed, even after I log out and quit, then launch localhost again, I am automatically logged in.

I am running again the same tools at the moment, making sure I named my history. If this problem (history disappearing) occurs again I will follow this up. The Galaxy version I am using is 20.05. I unfortunately did not record what it was before updating pip. Will do from now on.


1 Like

Thanks @Jean-Michel_Fustin_S

Do you remember if you “shut down” the local Galaxy before restarting your computer again and the other items? Details are probably captured in the logs so you can check there. If not shut down first, that may explain the problems.

Please check and we can follow up more from there. This isn’t expected behavior. If you are Ok redo-ing the analysis, that is your choice, but you probably want to make sure your installation is not corrupted in some way not obvious yet. Plus, get rid of the disconnected 500 Gb of data in a correct way. If that is needed – the data might be recoverable. I don’t know for certain myself but can recruit more admin help.