Jump to content

Shane

Moderators
  • Posts

    1035
  • Joined

  • Last visited

  • Days Won

    105

Posts posted by Shane

  1. You'll have to disable/uninstall DrivePool (or at least turn off all balancing) and run Recuva or whatever other tool you use on the individual drives. Don't forget to restore any recovered files to a completely different physical non-pool drive (not a drive that doesn't have any deleted-but-needs-to-be-recovered data).

    Then once you're satisfied you've got the files back, re-enable/install DrivePool.

  2. Correct, it doesn't support software RAID (because Windows insists on Dynamic disks for that). Hardware RAID is fine.

    All that matters to DrivePool itself, is that it can see the NTFS (or ReFS) volumes on Basic (not Dynamic) disks that are locally (not network) attached to the system.

    Doesn't matter whether that's via motherboard or expansion card, PCIE/SATA/USB/Thunderbolt/otherwise, in any combination.

     

  3. Hmm. You've got the "respect real-time file placement" ticked, and I was under the impression that the OFP plugin was real-time for new files. Try a Remeasure.

    But if it's not real-time, then you'd need to choose an Automatic balancing setting other than "Do not balance automatically" so it can be triggered.

    Edit: I've made a feature request to have the DrivePool GUI explicitily indicate which balancers are "real-time" and/or can force immediate balancing.

  4. Check that the ordered file placement plugin isn't being out-prioritised by another plugin that would prefer to write to the older hdd.  Also check the Balancing options (Manage Pool -> Balancing -> Settings), particularly the "Plug-in settings" and "File placement settings" sections.

  5. Unfortunately DrivePool doesn't support Dynamic disks (if I recall correctly it could be added but would require an overhaul of the balancing engine and further work on the duplication code), so it's not compatible with Windows' software RAID implementation.

    If your motherboard has even barebones RAID 0 support, that would suffice for DrivePool to "mirror" the pairs.

  6. As you've discovered you can manually move files at the individual drive level without immediate issues so long as the directory path is the same; the only potential issue is that DrivePool may then not have the correct measurements of the state of the pool for balancing; this can be remedied by manually forcing a re-measure (Manage Pool -> Remeasure) after you're done.

    Alternatively you could use Manage Pool -> Balancing -> File Placement to set rules for folders and/or files (e.g. "\tmp\today's project\*" or whatever) that you wish DrivePool to move to/from the NVMe drive; you could just (un)tick drives in the rule and then manually start a balancing pass to have it move files accordingly before and after your work.

  7. Hi! DrivePool only works with NTFS or ReFS formatted drives (there's also an advanced setting to allow both in the same pool but it's not recommended outside of migrating between the two formats); exFAT is not supported.

  8. I'll clarify, when I say recommendations I'm not asking for a spiel, just a "X is what I'm using" with an implied "it hasn't bitten me (yet)". Like how you mentioned which defragger you use.

    Also thankyou for highlighting the impact of sector scans re drive workload limits, and welcome to the forums!

  9. The problem is that SMART is not a strictly formalised standard; every manufacturer has their own implementation (sometimes differing even between models) using different scales and ranges and formats. Manufacturer A might use values between 100 and 199 at location 235 to measure an aspect of their drives while Manufacturer B might use values between 0 and 65535 at location 482 for that same aspect; and those values may not even be in the same scale (e.g. celsius vs kelvin vs fahrenheit for temperature). And so on ad nauseum.

    This means that if the particular SMART software you're using doesn't have a translation of whatever proprietary implementation it encounters sufficient to know that A=B means "bad" (and also track stuff like "not bad yet but it is getting worse"), it then has to rely entirely on whether (and when) the drive itself flags the results as "bad".

  10. Hmm. Tricky. I've done some testing; the SSD Optimizer balancer conflicts with the requirement to keep certain files on the SSDs because the Optimizer wants to flush the cache disks despite the File Placement rules (even if it does not do so on the first balancing run, it will do so on a subsequent run) regardless of the File Placement Settings in the Settings tab. I'm not sure if that's a bug or if it is working as intended... though I suspect maybe the former.

    I'll play around with it some more and see what I can come up with.

    EDIT: @Christopher (Drashna) is it possible to get the SSD Optimizer balancer to play nice with the File Placement rules? I've tried the various permutations possible in the Manage Pool -> Balancing -> Settings -> File placement settings section but even when it respects them on the first subsequent balancing pass it fails to do so on later balancing passes, i.e:

    1. I set up a pool with the SSDs marked as SSDs in the SSD Optimizer balancing plug-plugin.
    2. I create a "SSD" folder and set up a File Placement rule that any file in that folder should be kept on the SSDs.
    3. In File placement settings I tick "Balancing plug-ins respect the file placement rules".
    4. I copy files to the pool; they go into the SSD drives per the real-time requirement of the SSD Optimizer.
    5. I start a balancing pass; the File Placement rules are respected and the matching files stay on the SSDs while the rest are moved to the HDDs.
    6. I start another balancing pass; the File Placement rules are ignored and SSD Optimizer flushes all files on the SSDs to the HDDs.

    #6 shouldn't be happening? Doesn't seem to matter whether the other File placement settings are ticked or unticked.

×
×
  • Create New...