DougCube

Is there a way to view a log of past/dismissed notifications? If so, how?

This is probably obvious but I'm asking to make sure... While booted, I plugged in some new drives to preclear. Now that it's done, is it okay to unplug them while still booted? They are not part of the array of course. The SATA ports are AHCI. (I have a limited number of plugs and want to plug in other new drives to preclear without doing a reboot.) And when I'm ready to use these precleared drives in the future, will Unraid automatically recognize the preclear signatures still? Also, would it work on a different Unraid server than the one used to do the preclear? I've read similar questions but found nothing matching exactly what I described. Btw, are there any utilities in Windows to do a preclear that would make such a compatible signature? I mean without having to VM an Unraid within Windows. (I did see this asked in another thread, but that was 5 years old.) -Doug

While the server is up (for days), suddenly the cache drive disappears from the UI. But in the command line, the mount is still accessible, accessing files that I know are on the cache drive via SMB share still works, and running the Mover works fine. Both reads and writes to it seem to work fine. The drive does not appear in the Unassigned Devices section. I did not find anything relevant in the syslogs. I've seen this happen twice now. Rebooting the server fixes it but I'd like to prevent it from happening again. I suspect it may be happening when I update the Unassigned Devices plugin, but I can't be sure. Anyone else see this issue before? Any ideas what's going on here, or how to fix it?

Sounds like something went wrong with the preclear. Clicking the X should have no impact. Once the preclear is done, even before the post-read, there would be a signature written to the drive that Unraid picks up on when you later add it to the array so that it doesn't have to clear it again. I've never heard of this happening.

I am somewhat familiar with the Notify command, but how do I trigger one once Mover is done? Essentially, I'm looking for a way to have a script detect if the Mover is finished running. Or even better, have it trigger automatically each time. Speaking of which, wouldn't it have made sense for this feature to be built-in to Unraid?

Oh, good to know. I struck-out the incorrect info from my reply. Thanks. I might actually use this 'feature' going forward then. I just have a slight preference to group up drives of the same capacity and manufacturer.

The way the OP is worded makes me want to clarify that you can't have gaps in your array slot assignments. And yes, what you describe is considered "reordering" from the UnraidOS point of view (especially w.r.t. the parity2 calculation). And like others have said, the physical configuration doesn't matter at all -- you can change where a drive is physically mounted or which SATA/SAS cables/ports to use between boots and the OS will easily cope with it. (Although I wouldn't switch an array drive from being internally mounted to being in an external USB enclosure.) I used to be overly concerned with the ordering of the drives, but I eventually found that it's best to live with leaving the order alone. Although... I have requested a feature to be able to assign nicknames to the drives to make it easier to manage. But if you do choose to reorder, keep in mind that the 'included/excluded' drive selections in the shares do not automatically update to track the change -- you'll have to update those manually.

There is no limit on number of cache pools. Personally, I would go for the Pro license for future expandability, but I kept adding drives every few days until I hit 18 devices before my first trial key even expired. Currently, there can only be one 'Array' pool. Which conventional wisdom suggests should be all your HDDs together. In your case, the four large drives would be the 'Array' and you can use two of the SSDs to make a single btrfs RAID1 cache pool -- making it so you have redundancy as soon as the data writes to your cache. You can use remaining drives to make more cache pools, or be dedicated for VMs. And if you have even more SSDs to spare you could put them on the data array (a bit of a waste putting SSDs on the array IMO). Four HDDs is really just a start since you'll probably want to have one parity drive, which takes up nearly 25% of it. This sort of RAID5-ish setup is better suited for having more drives.

There seems to be a bug or oversight where hot-plugged drives do not give a preclear icon/option in the MAIN tab but it is still possible to run it from the TOOLS tab. (I tried obvious things like refreshing the page.) Please fix. Regarding the issue of this plugin disappearing after a reboot... After I updated to latest version to fix this very issue, I still needed to reinstall the plugin after a reboot -- at least twice more. (I should send you a diagnostics file again.)

To be clear, if I have a trial key that had expired 15 days ago, does getting a 15-day extension make any difference?

Does the 15-day extension start when the 30-day trial ends OR when you start the 15-day trial? (There's a subtle difference between the two.)

This question has been asked here many times. It can be done on a separate machine.

+1, useful feature

I was wrong earlier. But I meant not including parity, pool, boot, or unassigned. It is 28 right?

Other than the obvious (not being able to start the array), what features are blocked out when the trial key expires? I'm thinking of having a second Unraid machine that is dedicated to only running preclears, and prefer not to have to purchase another license for it. How about VMs? Does that feature work with expired trial key? I'm guessing dockers wouldn't work because they always need the array? Or can dockers be run from an unassigned device? And could you still do a disk share (SMB) of an unassigned device? I searched around and was unable to figure this stuff out -- or at least not explicitly spelled out anywhere.

Currently, if a share has "Included Disk(s)" or "Excluded Disk(s)" and you reorder the Disk slots, the selection for the shares are not updated accordingly. I sighted this in 6.9.0-beta30 but it's likely always been this way. What would be nice is if it did automatically update it for you. Even better is having a "reordering wizard", a "replace parity drive wizard", etc..

In "Array Devices" section the disk names are currently quite verbose "(<model number>_<serial number> - <size> (sb<letter>)". This is for the most part fine, but I would prefer to omit <size> since there is already a column showing it. The ability to rearrange and hide columns here would be nice too. I would also prefer to omit the <model number>. But even better might be to allow specifying your own identifier/nickname. And on top of that to give some formatting string, like "%nickname_%model_%serial (%sb)" -- you know, like printf().

+1 for multiple arrays Is this targeted for a specific future version yet? Like 7.0? Would this allow going past the limit of 24 disks?

I often run preclear on 3 to 4 at the same time. Saves a lot of time because it takes roughly the same amount of time as running just one.

For a large share, spanning several disks, using cache_pool=yes, allocation_method=fillup, and split_level=top_two... I don't know if this is due to Mover or the way I use the SMB or occasionally do manual 'rsync -avX' between disks, but when I check 'stat' on split directories, the modify timestamps are mismatching for some but identical for others. (I am only using user shares; never enabled disk shares.) 1. How is it intended to be maintained? (i.e. Should they ever differ? Is it okay for it to differ?) 2. Which modified date is given to the Windows SMB when they do differ? (first disk index takes priority?) 3. Is there some method/command to synchronize the modified_date across all parts of all split directories? I looked into using unBalance plugin but it doesn't run as root and I'm worried whether it deals with the above issues 'correctly' (whatever that may mean). Also I was worried about security because the unBalance page doesn't present a user login screen (does it use the same cookie?). I have since changed to split_level=manual but I know it only affects future writes to the share, it does not work retroactively and there is currently no built-in way for Unraid to re-adjust it. I wish there was tho. A way to tell Mover which disk to use for each directory, and for it to 'gather' for you.

Not sure how feasible this is, but when I have multiple browser windows/tabs open to the server, dismissing a notification in one does not dismiss it in all of them. Can this be made to dismiss in across all? I guess this might make sense to be user-specific even if logged in from multiple devices to have it synced, but a different 'user' should still see the notification. On top of this, if possible, in the agent API, to trigger notification dismissal there too.

Every time I reboot my server, the Preclear plugin disappears as if it uninstalled itself (or something else uninstalled it). And I have to reinstall it again. Is this a known issue? Anybody else seeing this issue? It's not happening with any of my other plugins. I'm on 6.9.0-beta30 with trial key.

This has probably been asked before... Is it possible to replace the parity disk with a new larger drive WITHOUT doing a data-rebuild? So it would copy from the old parity disk to the new disk and zero-out the rest without needing to spin-up the other disks. If possible, how? If not, shouldn't it be a requested feature? Forgive me for asking a basic question. I'm a newbie. But I did read https://wiki.unraid.net/The_parity_swap_procedure, which sort of shuffles the contents of 3 drives around. That is not quite what I want.

On my server, I have two NICs, each set to a different IP. Is there a way I can have the URLs setup so that it uses which ever IP it's currently under? Like a relative address instead of an absolute address if that makes sense. (The reason I'm needing this is because sometimes one connection goes down and I need to use the other.)