Rclone copy dry-run of big S3 bucket is taking super long time to start

[kapitainsky]

For comparison here you are example of well performing server - 80 million objects takes 2h to list

Rclone copy not copying

With --dump-headers, we can see that there were 80095 HTTP REQUEST, and as many HTTP RESPONSE (200). Processing took about 2 hours. The server has 30G of RAM. I restarted the copy with the options indicated, with the same result.