• Dana
    2
    The site has a 10 Mbps connection and I estimate it will take a week to upload a full backup.

    I created a Hybrid backup job with a monthly full backup to send the backups to Wasabi. Another job does a weekly full to a local drive. The problem is that the local backup job will not run while the Hybrid job is copying the backup to Wasabi. You would think that after the Hybrid job finishes making the image on the local storage that other jobs would be able to make backups. But no!

    I was also considering backing up to Wasabi but was not sure that the synthetic full would be significantly faster to reduce the time to a acceptable duration.

    Is it possible to use Cloudberry drive as a storage location? Would Cloudberry drive cache the changed files and send them to Wasabi?

    I have also considered creating splitting it in two jobs. An image job to create a local backup and then a file backup to copy the local backup to Wasabi. The only problem is that the backup in a backup would require two restores process to retrieve files.

    A scheme I have used is to backup to a NAS. Then RSYNC syncs the files to an offsite NAS. This works.

    I am testing a scheme of using a NAS to send the backup to Wasabi. However the backup is put into a folder in the bucket. So Cloudberry doesn't see the backup. Is there a way to quickly move blobs between folders so the backup could be put in the correct folder?

    If you have a scheme that can deal with a long upload time, I would like to hear it.

    Thanks
  • Anton Zorin
    30
    Hi ,
    Thanks for raising these questions.
    1. Cases with slow upload link are very common and we’ve built a short list of best practices:
    1.1. For image plan don’t use hybrid and use 2 plans instead: 1 for local and 1 for cloud.
    1.2. Use the synthetic full whenever possible (it’s supported for Wasabi, but not for hybrid plans) to re-use what’s already in the cloud and that would eliminate the need to upload everything from the local machine.
    1.3. Synthetic full will be much faster.
    2. Using Cloudberry/MSP360 Drive as a backup destination is a bad idea. Backup apps usually need high performance when executing their jobs, Drive is a personal use software that wasn’t designed to be a backup destination.
    3.
    The only problem is that the backup in a backup would require two restores process to retrieve files.
    can you please elaborate here? I’m not sure I understand it.
    4. Syncing a local backup folder to the backup folder in the cloud shouldn’t be performed manually. Currently the hybrid job does exactly this, but it doesn’t work together with the synthetic full, this is why it’s not recommended with slow links to the cloud. With the new backup format that we’re working on at the moment this will be addressed shortly after we release it.
    Thank you,
    Anton
  • Dana
    2
    For number 3 the scheme would create a local backup and then another backup job would backup the local storage to a cloud storage. Thus a backup inside a backup. Therefore, you have to restore the local backup and then restore the files.

    Anyway, I would like to try the synthetic backup since I am using Wasabi. But how does one do a seed for the initial backup? I have read the manual about how to use our high bandwidth to do the initial seed. However, it leaves out how the seed is created and then moved to our site for upload. It is left to speculation as to how to do this.

    My speculation is that a backup job is done to a USB drive on site. The USB drive is brought to our facility and then use storage explorer to copy the backup to Wasabi. Is that correct?
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal
youtube
tweet
Add a Comment