To generate the commands to abort multipart uploads to IBM cloud storage:
ibmcloud cos multipart-uploads --region us-south --output json --bucket filkarchive | jq -r '.Uploads[] | select(."Initiated" <= "2021-07-15") | "ibmcloud cos abort-multipart-upload --region us-south --bucket=filkarchive --key=\"\(.Key)\" --upload-id=\"\(.UploadId)\""'
You'll need the ibmcloud cli installed.
Set up an aws profile using aws configure --profile filkarchive
The Access Key ID and the Secret Access Keys come from cos_hmac_keys
in the IBMCloud Console Service credentials
Then, run the aws cli with
aws --endpoint-url https://s3.us-south.cloud-object-storage.appdomain.cloud/ --profile filkarchive [command]
To do anything with files in the cloud storage (download them, move them, rename them, etc.) they need to be restored from the archive tier. The ways I have found to do this are with the system console (very tedious), with the aws cli, or with Cyberduck.
For the aws cli, the command looks like:
aws --endpoint-url https://s3.us-south.cloud-object-storage.appdomain.cloud/ --profile filkarchive s3api restore-object --bucket filkarchive --restore-request 'Days=7' --key [filename]
From Cyberduck, you have to configure some Cyberduck hidden configuration options:
s3.glacier.restore.expiration.days=7
s3.glacier.restore.tier=Bulk
(You should be able to the expiration days to any reasonable value that works for you up to 30; I haven't tested whether any other restore tier works.)
Once you have these settings set, you can select "Restore" from the "File" menu, or from right-clicking on a file, and it will submit a restore request.
However you request it, the restore request takes up to 12 hours to complete the restoration from deep storage.