Transfer failed when attempting local export to external HD

On Kolibri server in W7 environment with over 150GB of content, I am having difficulty with exporting KA only. Other content (Tessa, Sikana, etc) exports fine. But when attempting to export any KA, I get error “Transfer failed. Please try again”. Upon trying again - same error. Regardless of how much of KA I select, even down to selecting one lesson of 487Kb, the same error.
Unfortunately, the KA is all I am needing to export for another installation. This server was built in an area with high speed internet and here we only have 3Mbps Vsat, so difficult to download 39GB of KA.
I have successfully copied the entire .kolibri folder to a new installation and it works, but would rather start with fresh database and add exported content, hence the need for exporting KA

Transfer failed screenshot

Exporting other content screenshot

Export succes of other than KA content

Hi @mrdavidhaag, this is Jonathan from the Learning Equality team. I’m sorry about your experience exporting KA; it’s odd that this is the only channel experiencing this issue.

If you can, could you share any error logs/messages that appear when you see the “Transfer failed” message. You should be able to find a “kolibri.log” file in the “.kolibri” location.

Could you also share the version of Kolibri you are running. Details about the external hard drive might be useful too.

Does the transfer fail immediately, or does it get to a certain percentage before failing?

As an experiment, could you try exporting single lessons of different types to the hard drive. Was the 487KB lesson a video, or an exercise? For example, try exporting a single video, the a single exercise.

Hi Jonathan, Thank you for the quick response.
I am unsure about error logs in the kolibri.log file (here on Drive) I see the export but unsure of corresponding errors.
Version is 0.11.0 in W7 OS
I tried to export a 96Kb quiz - no joy. Then tried 875Kb lesson (video) - same.
The export fails immediately, there is no progress of any percentage before failing.
I have tried to export to three different drives, two external and one internal. All with same result - only KA is having trouble with export. The rest of exporting is working normally and even large content like 44GB of Sikana exports successfully.

Thanks for the logs, there is some useful information in them.

One of the error messages sticks out to me, just as the export begins:

FileNotFoundError: [WinError 2] The system cannot find the file specified: ‘C:\Users\teacher\.kolibri\content\databases\1ceff53605e55bef987d88e0908658c5.sqlite3’

Would you be able to check that location to see if that file is there (it’s the database for Khan Academy)? Is ‘teacher’ the only user on the Kolibri server?

Yes, I saw that error and no that file does not exist

teacher is the only user in the W7 installation

There are several users in Kolibri. I just added myself and logged in and able to view KA content

It’s odd that the SQLite file for Khan Academy disappeared from your computer, but if you replace it, it will hopefully fix your problem. Here is a link to Version 14 of KA English from Kolibri Studio (which should be the same version as what you have currently, based on the logs):

If you replace it in .kolibri, the hope is that you won’t get this error anymore and you will be able to export to an external drive. Let us know if this works.

Thank you for the solution. It worked after a reset of the server.
I am downloading the KA Math content now.
The reason I am exporting the content is to setup a new Kolibri server for another installation. I wanted to be able to select the content necessary from a storage device rather than “bulk” moving the whole .kolibri folder to new server.
Question: Is there any disadvantage or possible trouble caused by moving the whole .kolibri folder from one server to another?

Ooops, sorry - spoke too soon.

Four hours later the Math section of KA was only 13% complete, so I cancelled.

I then tried to export MIT Blossoms and the prompt is endlessly spinning Preparing.

Please do not spend a lot of time on this issue because this server was built by a tech new to Kolibri and unable to get an Ubuntu machine running, so the tech opted for W7.

The whole reason I am using this server as source is that we only have 3Mbps Vsat connectivity here and getting the 155+ GB of content is not really feasible.

I believe I am going to start from scratch with new Ubuntu VM on a Proxmox server install 0.11.1 and then send it out to the coast where Fiber Optic internet is available.

I will try to delete the db file I put in and try again.

I am also concerned about using this server as source because it is W7 and all our other servers are Ubuntu. Plus this source appears to be corrupted or troubled in some sort of fashion.

Hence my previous question : Is there any disadvantage or possible trouble caused by moving the whole .kolibri folder from one server to another?

Thank you for your help.

This method is pretty safe, but it will transfer other kinds of data as well (users, classes, lessons, etc.). I think a technical advantage is that the operating system might be faster at transferring the data than through Kolibri, hence the slow progress you’re seeing on KA and MIT Blossoms.

So one thing you can try is to have a “clean” Kolibri server with no classes, extra users, and use that server’s .kolibri folder as a starting point for new servers you want to deploy.

I just learned from the team that there is kolibri manage deprovision command that you can use to “clean” a .kolibri folder of extra data, but leave the content intact.

Delete all facility user data from the local database, and put it back to a
clean state (but leaving content as-is).

Here is the output from the kolibri manage deprovision --help that explains how it’s used. In the most simple case, kolibri manage deprovision with no extra arguments should remove any non-content data.

usage: kolibri manage deprovision [-h] [--version] [-v {0,1,2,3}]
                                  [--settings SETTINGS]
                                  [--pythonpath PYTHONPATH] [--traceback]

Delete all facility user data from the local database, and put it back to a
clean state (but leaving content as-is).

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  -v {0,1,2,3}, --verbosity {0,1,2,3}
                        Verbosity level; 0=minimal output, 1=normal output,
                        2=verbose output, 3=very verbose output
  --settings SETTINGS   The Python path to a settings module, e.g.
                        "myproject.settings.main". If this isn't provided, the
                        DJANGO_SETTINGS_MODULE environment variable will be
  --pythonpath PYTHONPATH
                        A directory to add to the Python path, e.g.
  --traceback           Raise on CommandError exceptions
  --no-color            Don't colorize the command output.

Thank you for the expert tip from the team - good information to know.

I suppose if we had read the documentation fully we would already know this. I know from other projects I work on that a point of frustration for service providers like LE is the pretty common fact that technicians rarely fully read instructions before installation, commissioning and implementation. I often refer to documentation as “destruction manual” because it is not until technicians have destroyed something that we refer to the “instruction manual”.

The W7 source server is acting better now. After second reset, the Kolibri server took over 5 minutes to start. Not sure what took place in the long boot up, but exporting started working better.

I have checked two other servers that were built from this source and both do not have the KA database file. I plan to replace the missing KA database file in both servers and test.

Thank you to you and your team for excellent technical assistance and a robust / fruitful product.

1 Like

No problem. We will investigate why the KA database file might have disappeared, and might also add some documentation on the ‘deprovision’ command on our documentation It is not there currently, as this is a somewhat advanced feature.