Shane

Hi, if you check out https://wiki.covecube.com/StableBit_DrivePool_Advanced_Settings it describes how to edit advanced settings, one of which is CoveFs_WaitForVolumesOnMountMs to adjust how long DrivePool waits for Windows to mount the pool's disks. You could try increasing that value?

Hi Gabe, I'm seeing the same error. Does not occur with the previous release version 2.3.13.1687 when I revert so that could be a temporary workaround. I've opened a support ticket with StableBit.

Is it possible there are files in OneDrive's recycle bin online (checking via browser) that need to be deleted?

Hmm. Something does sound messed up there. If it helps, pools created by DrivePool show up as Volumes and Disks in Windows Disk Management (WDM). Clicking on the Pool volume in the volume list (default upper half) of WDM should highlight the related (virtual) disk in the graphical view (default bottom half) of WDM. So if everything was healthy you should have a number of Disks in WDM equal to the number of physical disks in your machine plus the number of active pools, starting from Disk 0. To start with I'd suggest powering off, pulling all three disks from the Orico, powering on, and checking WDM to see the results. Then repeat but with just the one suspicious disk put back in (D#1, not the others) and compare. The new disk in the second list should be that disk (though keep in mind that the number allocated to each disk may be different and the pool "disk" may also show up - it shouldn't, but it might). Then use DISKPART, LIST DISK, SELECT DISK # where # is the Disk number to select the suspicious disk and CLEAN to wipe that disk clean. Make sure you don't select the wrong disk! Then EXIT and use WDM (may need to use Action -> Rescan Disks first) to first initialise it as a GPT disk and second format it as a single NTFS volume. Then power down, put the other two good disks back in and see if that's sorted it.

Just to make sure I have this right, rephrased: For the old server: You had an old Win 10 machine with (at least) three physical disks, I'm going to call them D#1, D#2 and D#3. D#1 was the boot disk for the machine; it had the usual boot partitions/volumes (including one formatted as a NTFS volume, C drive) and either the C drive volume was part of the pool or there was a separate partition that was also formatted as a NTFS volume that was part of the pool; either way I'll call that volume V#1. D#2 and D#3 were each formatted as a single NTFS volume, which I'll call V#2 and V#3 respectively, and those volumes were also part of the above pool. During the upgrade: You first removed V#1 from the pool, then balanced the pool, then shut everything down. For the new server: The new machine is running Win 11 Pro (and presumably DrivePool is installed). You've connected the physical disks D#1, D#2 and D#3 from the old server to the new server via an Oricon external drive bay. You've then redone D#1 to just have a single, new, NTFS volume which I'll call V#1a. Is that all correct? At this point DrivePool should be able to see a pool consisting of V#2 and V#3 (on D#2 and D#3) as well as an available non-pooled spare to add, V#1a (on D#1), since you don't mention adding the new volume to the pool yet. However instead DrivePool can only see V#2 and V#3, and even Windows Disk Management cannot see V#1a (or even D#1?), and if you pull any one of D#1 thru D#3 the whole pool vanishes from the DrivePool GUI (when, I agree, the remaining disks should still be showing up)?

So basically something like the user being able to check a UI box for a given disk in the guest Scanner that says "Enable Passthrough via Local Network / StableBit Cloud+" and then if it can't retrieve SMART from that disk itself it'll first try querying its local peers for that disk and then if that doesn't work it'll try querying the StableBit Cloud+ service (if logged in)?

Hi wecpc, that procedure will work. It is preferable to connect all of a pool's drives at the same time (e.g. while Windows is not running or before DrivePool is installed) rather than one at a time so that DrivePool does not complain about missing drives, but DrivePool should automatically put any pool with missing disks into read-only mode until all disks are found. If you have any custom balancing/placement settings for your pool(s) you should take note of them before migrating as you may need to redo them on the new machine, but any duplication settings will be carried across automatically. All content should remain intact.

It will attempt to repeat the re-measuring when it next can; the measurements are used to help with balancing (though will still run even if no balancers are enabled). If you want to check whether DrivePool has finished measuring and (if any) balancing you can use the advanced settings "RunningFile_FileName" entry to have DrivePool place a specific temporary file somewhere while it is performing background task, and monitor that.

Hi sim_cubed, as far as I know it is as safe to power cycle a disk that's part of a pool as it is to power cycle a disk that isn't (e.g. if any new or changed files are being written to the pool at the time those may be lost). DrivePool's background balancing uses temporary files so existing content should not be at risk any more than they would be on a non-pool disk). When all disks are reconnected and the service is running DrivePool will re-measure the pool to check if any content has changed; if you've got a lot of files in the pool this can take a while.

That latest screenshot (after the night of balancing with only Scanner and Optimizer) looks promising even if having Equalizer as well in third place doesn't work out.

I would do so. Then just re-add it for Optimizer.

Hi! I would first order the priority of balancers as follows: StableBit Scanner - if a disk goes bad you want the evacuation to take priority over anything else SSD Optimizer - when disks aren't bad you want the SSD caching to happen before anything else Disk Space Equalizer - then you want the archive disks equalized (per the manual's notes this balancer overrides anything beneath it) I also seem to recall that in some cases the Equalizer and Optimizer do not play nice together - both are "aggressive" - so of the two you might try enabling only the Equaliser first and then once the HDDs are balanced to your liking enabling the Optimizer instead (new files by default should go to whichever drive has the most available space, so if they're all the same capacity over time they should stay more or less evenly filled).

I don't actually know the exact algorithm it uses; with three drives and the goal of maximising free space on multiple drives to fit as many duplicated (x2) files as possible I'd expect it should (attempt to) end up with one drive full and the other two with a roughly equal amount of free space. For example if you have drives A____ B____ and filled them with x2 duplicated files 1 2 3 4 then you'd end up with A1234 B1234. If you then added drive C____ the balancer could try to arrange them something like A1234 B12__ C34__ so it'd then have the most room for additional x2 duplicated files. Something I could play around with to test if I had more minutes of free time lately.

Huh. Fair, that'd be weird. Any chance it could have gotten a start on scanning the sectors before it had to pause for whatever it's waiting for and you saw its (lack of) progress? Or is the screenshot from like "thirty seconds after being plugged in for the first time" and it couldn't possibly have made that much progress even if it started scanning straight away?

A4. Green blocks are groups of sectors that Scanner has already checked and found to be readable (they may initially start off blue as Scanner works). Grey blocks are those that Scanner has never checked yet (or which you've manually cleared to be re-read). If it seems to be stuck on "Waiting to Scan" then either you've turned off automatic scanning or Scanner's settings are having it wait for something else to finish with the disk (Scanner has a LOT of settings as to when and why to scan or to not scan, which can be accessed via Settings -> Scanner Settings or you can select a "set" of defaults to use via Settings -> Quick Settings). Also if you minimise the File system health section for J:\ in that screenshot you should see the Disk sectors section show an additional button in the left column of options, which (unless already scanning that disk) should be a blue button with a white triangle, which can be used to manually start a scan of that disk.

I suppose I can't rule out a haunted port, but I'd still place my monopoly money on door number two (drive belatedly self-corrected). Wondering, if you clear the blocks and rescan it does Scanner still find the same damage?

Hi Daxar. If you add a third drive but keep the duplication at x2, which drives any given file ends up on will depend on how you have the balancers set. The default set of balancers includes the Duplication Space Optimizer; its job is to optimize the pool to make the most use of your free spacefor duplicating files, so it will want to move some of the existing files to the third drive to free up space for more duplicated files.

Basically SMART errors can be divided into "it may fix itself" and "it won't fix itself". Neither is necessarily "replace it ASAP" but it's more probable with the latter. Seek error rate is with respect to how well the drive is moving the head(s) to the correct position and can change over time due to various factors (vibration, temperature, wear, etc). It is not usually by itself a sign of a failing drive (e.g. a consistently high seek error rate can indicate a defect that can't be compensated for but isn't ultimately preventing the drive from reading and writing data, just slowing it down). Drives can self-correct to an extent; if a SMART error pops up and then disappears shortly after and doesn't return, the drive's own logic has most likely taken care of it. E.g. in the case of seek errors, the drive may adjust parameters of the head movements to compensate. Scanner polls SMART frequently by default and thus it can often catch these temporary errors before a drive can finish self-correcting the issue. These temporary errors can also be due to circumstances beyond a drive's control (e.g. occasionally running hot due to high load or insufficient cooling, or something in the chassis is loose and causing vibrations) so it's still useful to know about. P.S. DrivePool's Scanner balancer plug-in has two tiers of whether to automatically evacuate, warnings (basically SMART has reported something to complain about) and damage (SMART is reporting there's something the drive can't fix or Scanner's own scans are finding damage).

Yes. I know at one point Chris mentions their pool had 26 drives in it, that was a while back. Glancing through the same thread another user's pool is 35 drives for example. I'm pretty sure I've seen screenshots of pools with ridiculous numbers of drives (40+) elsewhere in the forums but I can't remember where or exact numbers offhand. Theoretically DP can handle as many drives as Windows can support; the real limit is having good hardware. P.S. The trick with hardware RAID, zfs et al is you don't make one giant array, you make an array of smaller (three to six disks each) arrays. The pro is you can get things like parity protection, bitrot healing, stuff like that. The con is the extra micromanagement and how fiddly adding and removing disks can get. If you have a machine with good hardware RAID management you can have DrivePool do the pooling for all the small arrays so you can get the best of both worlds, but it's still something I'd only recommend planning out carefully.

If the drive is sufficiently bad, it may not be removable via the UI options while it's still physically connected to the machine. If you have duplication enabled for its content you can still physically remove it and then remove it in the UI, and DrivePool will reduplicate the content from and to the remaining disks. If you don't have duplication enabled for its content and don't want to restore from a backup you could manually copy that content to a spare or new disk (how well this works will of course depend on how it's going bad) before physically removing it so that you can reinsert the content back into the pool afterwards.

I would suggest opening a ticket with StableBit with what's happened and your workaround, hopefully they can both resolve it permanently for you and fix whatever the underlying cause is. Tagging @Christopher (Drashna) in the meantime.

Hmm. What happens if you run "dpcmd unignore-poolpart poolpartfullpath" for one of the "missing" disks? For example if you mounted one of the "missing" disks under the drive letter "M" you'd run something like (replacing the poolpart's folder name appropriately): dpcmd unignore-poolpart m:\poolpart.12345678-90ab-cdef-1234-567890abcdef Or if you had all of your poolpart disks mounted as folders under a physical disk instead of having drive letters of their own, it'd be something like: dpcmd unignore-poolpart e:\mounts\disk05\poolpart.12345678-90ab-cdef-1234-567890abcdef Does that disk come back? Just wondering if the unexpected restart in the middle of adding the new disks mangled the pool's part tagging.

The scanner plugin should since "Unless the drive is being emptied" is still ticked. The optimizer plugin I'm unsure of.

A1. It looks like it's based on the oldest scanned block. If you expand the drive (the + sign) it shows the sector scanning and you can see the date and time of when that happened to the upper right of the grid map, which at least on mine matches up. A2. The tooltip mentions that it's "only when disks have to be re-checked" and the tooltip for "unless disk scans are past due by" mentions it won't do anything until the disk has already been scanned at least once? A3. Could be Windows-related; I've had Windows insist a drive is in use even when Resource Monitor is saying there aren't any files open? Welcome to DrivePool and Scanner!

I think un-ticking "Balancing -> Settings -> File placement rules respect real-time file placement limits set by the balancing plug-ins." is what's needed? You want the VMs to be sorted into the 4TB SSD by the File placement rules more than you want them to be cached into and emptied out of the 1TB SSD by the SSD Optimizer plug-in, and perhaps without that un-ticking the SSD Optimizer might be ignoring the File placement rules when it's emptying the "SSD" disk into the "Archive" disks?