• Tylan
    2
    I'm a MBS User / MSP. My test machine is W10, imaged based backups to a local hdd, and then it's chained to a imaged based job to WHS. The WHS job isn't scheduled except for the monthly full synthetic.

    I have a few questions:

    How can I tell a wasabi synthetic full backup from a block level? My office machine has one full, and a ton of block level jobs... Like 6 months of block level. I really thought the syn full would look like a full. Is the keep the last version setting causing me to have to have issues here?

    Should I assume that storing an image based backup with 3 month retention will equal 3x my machine? It seems to be more than that. What should I set the job at to have 90 days worth of files? I assume that makes the most sense since I have to pay for 90 days with Wasabi once it's deleted. It looks like 3 month retention makes you store 4 full backups + incrementals?

    Third, will a synthetic full backup be the same size in the cloud as a full backup... Or does it somehow reference the original full? I read it as it copies the original full + the new incremental = new syn full. That makes me think each syn full will take up the same space as a full.

    Thanks!
  • Matt
    91
    In most of the cases simply specifying full backup for the second plan in chain should help, but make sure you have a a pretty big Windows between the runs to make sure the software is able to finish one full backup and start the next one. Basically, the second plan only needs to have full backup scheduled and that's it.

    Now, regarding the rest of your questions:
    Synthetic full will appear just like a regular full, so you don't need to worry about that. As for 6 months of block-level backups, this actually might be a problem. A small one, but a problem nonetheless.
    The thing is, purging only happens upon full backups and only in chains of full+block, and if you're setting up a time-based retention purge of the oldest chain will happen only when its last block is old enough(let's say 90 days), meaning that previous backups will be much older and they'll still will occupy your storage space.

    My suggestions on setting up your retention policy:
    Since synthetic full will take much less time to complete I recommend running full backup once a week and use your existing retention settings. That should be optimal in terms of purging, versioning, used space and the amount of data kept on storage side without triggering early deletion fees from Wasabi.

    Regarding full backup size: It'll be the size of a proper full backup, but the software will reference previous full + blocks to merge all previously uploaded data with changes that have been performed since the last plan run. Synthetic full is usually about 5 times faster than regular full because of that. On some configurations I got even better results.
  • Tylan
    2
    Awesome info. I think I figured it out not long after posting the original message.

    Full local image: First Sunday at 11:00PM
    Daily local image: 11:00pm
    Wasabi image: chained to previous job
    Wasabi full: Second Sunday at 11:00 PM

    As far as I can tell the Wasabi full syn has never run! I don't remember seeing it as a missed job. I wonder if that's because the job itself is unscheduled except for the full? Either I missed it or I wasn't properly notified. i'm going to have to see if I can find old emails. For some reason I thought the second image job would queue up. Would you think notifications work properly for that chained job? Ugh, this wouldn't have happened if we had hybrids (w/ syn) as an option.

    So, why run the Wasabi full image weekly instead of monthly? Also, I don't really need 90 days of backups, but I wanted to do whatever makes most sense because wasabi charges for 90 days whether you keep it or delete it.
  • Matt
    91
    Most likely the full backup was in overdue status but then block backups cleared it out. Try running your backups with the setup you mentioned, should work without problems.

    I usually recommend running fulls every week because it'll make versioning and purging more consistent. If additional versions on storage side is not a problem then you can run full once a month. In your case that's pretty sensible, since Wasabi charges for early deletions.
  • Tylan
    2
    Are you saying that the syn full job scheduled for 11 PM should work? It does not. I see errors about it not being able to run because another image job is already running. I think I have to choose a different time slot for it.
  • Matt
    91
    If it still doesn't run send us the logs via tools > diagnostic menu. Make sure that skipped full backup is (supposedly) among the last 10 runs, that way we get more info on that.
  • Tylan
    2
    I'm saying the full syn doesn't run because the local daily block level job in the chain is running. I think it will work as long as I schedule the full syn for a different time than the nightly local image.
  • Tylan
    2
    Just to update here...
    Modified the timing on my test system. Full syn works when there isn't a scheduling conflict. There needs to be a proper queuing method here or something that gives full backups higher priority than block levels. Trying to guess when jobs will finish is a pain. Maybe the easy solution is to schedule the monthly full syn to run 5 minutes before the daily job. That would at least make the less important job get skipped vs the full. I still feel like I lacked proper notification of the missed full job,
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal
youtube
tweet
Add a Comment