How to seed a backup to Minio S3?


#1

Hello,

I have set up Minio on my server in JBOD mode and want to use it as an S3 compatible backup endpoint for Duplicati running on my other computers. How can I pre-seed this destination? I tried backing up to a local drive first, then copying the files into the Minio data folders on the drives of the server. However when I log in to the Minio web browser I don’t see the Duplicati backup files there. (Although it does seem to be aware of them as the Space Used shows 300+ GB).

Is there another way I should be seeding this destination before pointing Duplicati at it to reconnect the backup job?

Much thanks for any help anyone can provide!


#2

Hi @jaydee. You seem to doing it correctly although I’m not quite sure why you’re not seeing the data you seeded. Have you tried re-configuring your job with the new storage destination?


#3

You may just be able to drop it into the interface.

If it’s setup to split up the data or add checksums, then it will need to add metadata for all the uploaded files so you can’t directly put stuff into it’s data directory.


#4

Thanks very much @Pectojin and @samw! Given your comments on checksums, I kind of gave up on this and just did the backup using minio as the destination in duplicati. It took a while but now the backup files show properly in minio. Now I’ve run into need for this again. I want to set up the same backup job to go to a remote minio server. The difference is the remote server only has one disk while the local one has the minio data spread across six disks in erasure mode. I’m wondering, can I just copy all of the duplicati files from the six disks to a portable drive and then copy them all combined into the data directory on single disk on the remote server then connect the duplicati backup job to it? Will that work? Thanks again!


#6

Are you copy/ pasting in minio data folder??

U need to copy your duplicati data with either mc or rclone to your MINIO instance.


#7

Jaydee, Minio is pretty funny about how data is moved into it. As @saviodsouza mentions, you really need to use the provided MC tool or something like cyberduck/S3 explorer. Moreover, incomplete transfers (as of the last time I used minio) will sit in a hidden cache that takes up space, but cannot be viewed. There is a process for clearing this cache manually, though finding the instructions for doing so may be a pain. Really, seeding it via the actual backup is the most reliable way, and IMHO the way you should be doing it considering minio’s quirky behavior.