• dbdata
    I have a Server 2012 with Cloudberry backup to Amazon. The backup is set for block level, Real Time backup.

    To me "real time" would mean that whenever a file changes, that file is backed up. Further guessing, I'm assuming that every file system has some sort of timestamp, changed when a file changes, that CB could look at, see that the "R:" filesystem has been updated as of {timestamp} and then go searching for files.

    That's my guess.

    But periodically I log on and see "real time backup starting in xxx seconds" and my inner moron says that if it happens periodically, then it's batch and not real time. On the other hand maybe it's best that I don't care.

    But here is my other issue:

    I have a SHARE on the server (the S: drive - not very imaginative) where another system places three files nightly. The files are (example only) A.gz (30gb) AB.gz (8gb) and XR.gb (22gb)

    So what I mean is that every night, another system replaces the previous night's files.

    And Cloudberry backup doesn't seem to notice and back them up. The timestamp under windows (navigate to S:\storage\) shows the newer date & time ... but they don't get backed up unless I manually run the backup.

    So I'd like someone to tell me what to do
    (run screaming into traffic has already been suggested and discarded)
  • David Gugick
    Schedule your backups instead of using real-time. Real-time cannot monitor shares and under stress (when many files are changing quickly) it may even miss files. It's much better to schedule your backups as needed, even if it's something like every hour.
  • dbdata

    OK --to be clear the share is a local filesystem to the server - but I will take tour advice

    what is your guess about block-level in this instance?
    If a.gz is a toally new file from the one from yesterday ... but the first 70% of the file is identical, will it see it AS identical?
  • David Gugick
    Block-level will identify the differences in the files and only back up the changes. The full backup you'll have to schedule to run with some frequency will only back up the files in full that were backed up using block-level backups. That's done so we can properly manage retention as set in the backup plan.
  • David Gugick
    Having said that, a full will run instead of block-level if the percent of file changes are over your set limit. See this post for a full explanation: https://www.cloudberrylab.com/resources/blog/block-level-backup/
Add a Comment