I had to delete an AWS S3 bucket with a ton of stuff in it. One bucket with millions of files. It was for work and my work laptop was very locked down. I had the aws client and python. Now I know that there are better ways but I was busy and did not have a lot of time to devote to it. So issues came up. I did not know the files were versioned. I kept getting ssl-warnings from aws cli and python kept throwing ssl-warnings. Finally, I would be kicked off every hour because sessions to AWS was limited to 1 hour.
I had to stop getting kicked off. So I had to renew my session before the 1 hour limitation was up. I did this with a python script and put it in a for loop that slept for 45 minutes and then renewed my session. You can find info on getting session tokens at AWS id credentials.
I was going to delete all files in my bucket. Easy, I thought, just do an aws cli remove bucket command. However, I got the dreaded ssl-warnings from the aws cli command. To prevent this put the –no-verify-ssl argument at the end of the command
aws s3 rb s3://mybucket-name --force --no-verify-ssl
The ‘–force’ removes all file and then removes the bucket. After many hours it finished but did not delete the bucket. The issue I had was versioned files in the bucket. There is a way with aws cli but it was easier to use python. If you get python errors with unverified HTTPS request, you can squash them with
export PYTHONWARNINGS="ignore:Unverified HTTPS request"
The script to delete versioned objects I used is
BUCKET = 'mybucket-name'
s3 = boto3.resource('s3',verify=False)
bucket = s3.Bucket(BUCKET)
If you don’t use boto3 for AWS python development you are putting to much effort in, use boto3. I did not want the AWS ssl verified, so that is why I put int ‘verify=False’. After several more hours the versioned files were deleted and I could remove the bucket.