Hi,
I'm running on Dell ECS storage appliance, which offers S3 compatibility.
yes http://ECS.ai is the DNS name of the server.
ClearML-models is the bucket.
Let me try with ip:port.
Hi SubstantialElk6 , I have a few questions:
Are you running minio or some other service to store the data? Is http://ecs.ai the url of the server and clearml-models the bucket? Can you check if accessing by ip:port instead of url works?
Hi TimelyPenguin76 ,
If you notice in the last screenshot, it state the bucket name to be http://ecs.ai . It then it tries to open http://s3.amazonaws.com/ecs.ai/clearml-models/artifact/uploading_file?X-Amz-Algorithm= ....
No, i can't see the files. But i can see if i don't use ':port' in the URL when uploading. I can't access the machine today, i'll try to check the S3 logs when i'm back.
Hi SubstantialElk6 , was this behaviour different in previous versions? If so, in which version?
Hi SubstantialElk6 ,
Can you add a screenshot of it? what do you have as MODEL URL ?
Hi, when i tried ip:port, it references the right host and bucket....BUT... the file is not found on the ECS S3 even though i can see from the logs that it states Completed model upload to s3://ecs.ai:80/clearml-models/artifacts/ ...
Hi. Yup the model was not physically uploaded with the up:port into the bucket, although ClearML does indicate that it's there, except that I can't download it. I also verified this with another S3 client, the model was not there as well.
SubstantialElk6 - can you please check the S3 configuration in your clearml.conf? Make sure if has the following:host: "ecs.ai:80" key: "your key" secret: "your secret" multipart: false secure: false
SubstantialElk6 - Not sure I understood. When you set output_uri=' s3://ecs.ai:80/clearml-models/artifact ' , is the model not uploaded to the correct bucket?
I didn't track the version on this change in behaviour. But last I tried it was able to download the content after I provide the credentials.
I'm asking since as far as I know there was no change in the WebApp S3 driver from previous versions...
This is strange then, is it possible for clearml logs to register successfully saving into a S3 storage when actually it isn't? For example, i've seen in past experiences with certain S3 client that saved onto a local folder called 's3:/' instead of putting it on S3 storage itself.
Well, a big is always possible, but we haven't seen that kind of behavior reported. As for storing in another place, that's only if the prefix somehow starts with a slash, and we know it doesn't since you can see that in the registered url.
The port (:80 in your case) is required, since otherwise both SDK and UI assume this is an AWS S3 bucket
Can you not see the files using the ECS files browser?