ServerError: server failed with HTTP status code 503

Question

ServerError: server failed with HTTP status code 503

Closed this issue 9 months ago · 9 comments

When I tried to download data using the sample code, I received a 503 error response.

import openml
# List all datasets and their properties
openml.datasets.list_datasets(output_format="dataframe")
# Get dataset by ID
dataset = openml.datasets.get_dataset(61)
# Get dataset by name
dataset = openml.datasets.get_dataset('Fashion-MNIST')
# Get the data itself as a dataframe (or otherwise)
X, y, _, _ = dataset.get_data(dataset_format="dataframe")

I get：
ServerError: server failed with HTTP status code 503

Answer 1 · 2023-07-19T04:40:25.000Z

Same issue here with another dataset

Answer 2 · 2023-07-19T05:53:19.000Z

Good morning, Thanks for the error report. Yesterday our server maintenance led to an issue. We're working on fixing it, but the expectation is that this will take some time. Best, Jos

…

________________________________ From: Charles Beauville ***@***.***> Sent: Wednesday, July 19, 2023 06:40 To: openml/OpenML ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [openml/OpenML] ServerError: server failed with HTTP status code 503 (Issue #1196) Same issue here with another dataset — Reply to this email directly, view it on GitHub<#1196 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABPE6ONKPGYZZ4YE4I5EWBLXQ5QMLANCNFSM6AAAAAA2PIYSTA>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

Answer 3 · 2023-07-19T05:59:59.000Z

I'm delighted to hear that.
Could you please inform us once the issue has been successfully resolved?
We would greatly appreciate it.

Answer 4 · 2023-07-20T09:52:56.000Z

Having the same issue here.

Answer 5 · 2023-07-23T22:40:45.000Z

Good morning, Thanks for the error report. Yesterday our server maintenance led to an issue. We're working on fixing it, but the expectation is that this will take some time. Best, Jos
…
________________________________ From: Charles Beauville @.> Sent: Wednesday, July 19, 2023 06:40 To: openml/OpenML @.> Cc: Subscribed @.> Subject: Re: [openml/OpenML] ServerError: server failed with HTTP status code 503 (Issue #1196) Same issue here with another dataset — Reply to this email directly, view it on GitHub<#1196 (comment)>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABPE6ONKPGYZZ4YE4I5EWBLXQ5QMLANCNFSM6AAAAAA2PIYSTA. You are receiving this because you are subscribed to this thread.Message ID: @.>

Any updates ??

Answer 6 · 2023-07-24T08:47:02.000Z

Hi all, Unfortunately, we're still facing server issues. Due to being behind on maintenance, a small change lead to a series of issues from which we are still recovering. I'm hopeful though. We seem to have arrived at problems in the last parts of the components. There's a fair change that we fix it today/tomorrow. It's difficult to predict though - new problems can pop up. For the users of the Python package: We pushed a new release for our python package which allows it to fall back on the arff files correctly. It will be a bit slower and memory intensive loading datasets that way, but for the most part it should work. So update your openml python to 0.14.1! As soon as everything is fixed (hopefully soon), we'll send a new update. Best, Jos

…

________________________________ From: nana247 ***@***.***> Sent: Monday, July 24, 2023 00:40 To: openml/OpenML ***@***.***> Cc: Velde, Jos van der ***@***.***>; Comment ***@***.***> Subject: Re: [openml/OpenML] ServerError: server failed with HTTP status code 503 (Issue #1196) Good morning, Thanks for the error report. Yesterday our server maintenance led to an issue. We're working on fixing it, but the expectation is that this will take some time. Best, Jos …

________________________________ From: Charles Beauville @.> Sent: Wednesday, July 19, 2023 06:40 To: openml/OpenML @.> Cc: Subscribed @.> Subject: Re: [openml/OpenML] ServerError: server failed with HTTP status code 503 (Issue #1196<#1196>) Same issue here with another dataset — Reply to this email directly, view it on GitHub<#1196 (comment)<#1196 (comment)>>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABPE6ONKPGYZZ4YE4I5EWBLXQ5QMLANCNFSM6AAAAAA2PIYSTA. You are receiving this because you are subscribed to this thread.Message ID: @.> Any updates ?? — Reply to this email directly, view it on GitHub<#1196 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABPE6OMZVLLDFUVQQ2UXKTDXRWR7RANCNFSM6AAAAAA2PIYSTA>. You are receiving this because you commented.Message ID: ***@***.***>

Answer 7 · 2023-07-24T09:09:12.000Z

@josvandervelde Thanks for the temporary fix!

Answer 8 · 2023-07-27T09:05:00.000Z

Hi everyone, We can finally share good news: the parquet problems are fixed! If you experience errors such as "Parquet magic bytes not found in footer. Either the file is corrupted or this is not a parquet file.," you'll have to remove your cache (on default on linux the cache is at ~/.cache/openml/org/openml/www/datasets). If you experience any other errors, please let us know. It took us longer than expected - one problem lead to another. There were many "hidden" problems due to being behind on maintenance. Now our infra is updated, so we expect never to face so much downtime again for the parquet files. So there's a silver lining after all. Best, Jos

…

________________________________ From: Charles Beauville ***@***.***> Sent: Monday, July 24, 2023 11:09 To: openml/OpenML ***@***.***> Cc: Velde, Jos van der ***@***.***>; Mention ***@***.***> Subject: Re: [openml/OpenML] ServerError: server failed with HTTP status code 503 (Issue #1196) @josvandervelde<https://github.com/josvandervelde> Thanks for the temporary fix! — Reply to this email directly, view it on GitHub<#1196 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABPE6OPOPPIEL3DTX4FA44TXRY3UHANCNFSM6AAAAAA2PIYSTA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

Answer 9 · 2023-11-01T00:16:17.000Z

Closing because this was fixed a while ago.