Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

WIP data-mover #2474

Open
wants to merge 9 commits into
base: master
Choose a base branch
from
Open

WIP data-mover #2474

wants to merge 9 commits into from

Conversation

jlento
Copy link
Collaborator

@jlento jlento commented Mar 5, 2025

@jlento
Copy link
Collaborator Author

jlento commented Mar 5, 2025

Initial version, both documentation and the actual tool WIP. Perfectly good for testing and feedback, though. Thank you :D

@jlento jlento requested a review from amsaren March 6, 2025 07:40
@jlento
Copy link
Collaborator Author

jlento commented Mar 6, 2025

Hola Dear Reviewers! I added plenty of you, and I'm happy to add more. The idea is that if you have time, you can also test the tool and try to break it :) Also, all the feedback on ease-of-use, how your tests worked out, etc is more than welcome.

https://gitlab.ci.csc.fi/compen/hpc-environment/data-mover/-/issues
https://rc.csc.fi/channel/Data-tiering-tool

than would be practical in regular interactive login sessions.

Below is a guide for a simple scenario, moving data from Puhti project scratch
directory to corresponding project in Allas, and then back. Similar works with
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

similar what works? Unclear

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess you mean exact same instructions work on mahti, which is true. Lumi-O is slightly different, need to specify more

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These object storages are really bad from the perspective of traditional HPC use. The mapping between filesystem to object storage is far from 1-to-1, object storage is completely separate machine with it's own authentication and authorisation, there are many different transfer tools/clients, APIs, and object storage server configurations, all different and often incompatible, instead of OS just handling it... I started writing it all out, noticed that it would be a long article, wrote TLDR text (what it is now), and deleted the start of the more complete guide. This tool is supposed to be easy to use. If the documentation is long, it means the tool is not easy to use :)

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll see if I can redirect the reader quicker to more comprehensive docs for using other services than puhti and allas

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok. I reread the doc. Similar is exactly how it is. Very unclear, but truthfully so :)

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you could just say that easy instructions work the same way in Mahti. Cross service usage, using Lumi-O instead of allas is possible, please go read the advanced section if you are interested


## Moving data from Puhti to Allas

1. Put the data in a single directory, for example
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Interesting wording, I always thougth that users would take an existing folder and move that. Of course, one could also put data into a new one. Perhaps could be clarified though

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good point. I think I have an idea how to fix that.

```
dm import /scratch/project_<projid>/exampledir
```

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What happens for the overlapping files in exampledir?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How do you remove old exports?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What do you mean by overlapping files?

Deleting an export from Allas means you moved something that you should have simply deleted in the first place :D

Ok, there is dm delete ..., says dm help :D There are also lot's of other dm subcommands, but my concern is to make this documentation short. If it is long, it looks like the tool is complicated and not easy to use.

[Puhti web interface](https://puhti.csc.fi). Open "Cloud storage configuration" from the
"Tools" drop-down menu, and
[create Allas S3 rclone configuration for the project](../../computing/webinterface/file-browser.md#accessing-allas-and-lumi-o).

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok. So this uses the the Open on demand style s3 configuration. This will be bit confusing for the old users of S3 allas.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants