Maintenance work November 24-28

[joaquinvanschoren]

In the week of 24-28 November, we're planning major maintenance work on our server infrastructure, moving fully to a K8s infrastructure and upgrading hardware. During this time, we'll have to put the production server into read-only mode. However, we'll do our best to make sure that the impact is minimal.

Current timeline:

• Monday 24 November, 9:00 CET: Switch openml.org to a read-only server.
• Friday 28 November, 12:00 CET: Switch back to full production server.

At the times of these switches, there may be temporary downtime.

We'll try to make the impact as minimal as possible, and the test server will remain available for development.

If you have further questions, please discuss below. Thanks for your understanding!

[joaquinvanschoren]

Update 25/11, 12:30: We had to redeploy the server, the search engine is repopulating, this might take a little while.

[joaquinvanschoren]

We started switching back to our production server. There may be interruptions due to DNS updates.

[PGijsbers]

We have now switched back to the production servers, but not all systems are operational yet.
If you encounter any issue, please let us know here.

Known issues:

• It is currently not possible to create a new user account or reset your password, as they rely on an e-mail server.
• Uploaded data will not be processed.
• It is currently not possible to log in.
• Runs can currently not be downloaded
• Using abbreviated urls leads to an internal server error (e.g., https://openml.org/d/61)

We expect to be able to gradually resolve these issues over the following week.

[josvandervelde]

Update: the first known issue (sign up / reset password) has been solved

[loremss3105-eng]

Good morning, today is december 6 and the server is not working, do you have more information regarding the issue?

[joaquinvanschoren]

Apologies, we were having some intermittent server issues as we are still working on improving the infrastructure. At the moment things seem to be stable again. We hope to finalize the last aspects of this maintenance as soon as possible.

[merveguel]

Hi team, signup page seems still down (since late December) - any chance for an update?

[josvandervelde]

Hi, the signup page is working again!

[rossGardiner]

Hi, I noticed that some datasets are not available.

E.g. if I run:

openml_id = 44306
ds = openml.datasets.get_dataset(openml_id, download_data=True, download_all_files=True)

error message:

WARNING:openml.datasets.functions:Could not download file from https://openml.org/datasets/0004/44306/dataset_44306.pq: HTTPSConnectionPool(host='www.openml.org', port=443): Max retries exceeded with url: https://www.openml.org/datasets/0004/44306/dataset_44306.pq (Caused by ResponseError('too many redirects'))
WARNING:openml.datasets.functions:Failed to download parquet, fallback on ARFF.

I cannot access the images.

Any advice or updates appreciated. Thank you!

[PGijsbers]

Thanks for the report, we'll look into this.

[PGijsbers]

Should be fixed. Please clear your cache before retrying. You can find the cache location with python -c "import openml; print(openml.config.get_cache_directory())". Specifically only removing the xml file from the subdirectory of the dataset (e.g,. org/openml/www/datasets/44306) should be sufficient.

[rossGardiner]

Fixed, thank you.

[Manas-7854]

Hi I am not being able to generate a test api key, its showing loading... for a while now. Is this related to the maintenance?