PPH

Upgraded from 6.12.13 to 6.12.14 - no issues so far.

Thank you for the detailed response. 👍

Hi @mgutt, I can see from your various screenshots that you have chosen to use a traditional unRAID Array using (SATA) SSDs. I thought SSDs within the unRAID Array was not fully supported (issues with TRIM?). Do you know if these are no longer valid and SSDs are now suitable for being used within a traditional unRAID Array? Thanks.

Successfully upgraded my "virtual" testing unRAID instance from 6.12.12 -> 6.12.13 and my "Media Server" instance from 6.12.11 -> 6.12.13 . Happy to report 'disk spin down' now working on both unRAID Servers.

Upgraded my "virtual" testing unRAID instance from 6.12.11 to 6.12.12. - no issues upgrading but have the same issue as others whereby no disks will spin down.

With my recommendation of configuring a 'traditional' unRAID "Array" (HDD) with a mirrored "Pool" (NVMe) and a "media" Share - files downloaded to the "media" share will be stored on the NVMe disks and then when "mover" is run (typically once a day) the files will be 'moved' from the NVMe "Pool" to the HDD "Array". With this type of configuration you get the speed/benefit of the NVMe disks with the storage capacity of the HDDs. "Mover" (Settings -> Scheduler) can be adjusted to run on different schedules and there is also a "Mover Tuning" plug-in that offers many options on which files should be moved (although, with being new to unRAID, I would stick with the standard "Mover" until more familiar with unRAID). Note that files are only moved between different storage locations if the "Shares" settings have both the "Primary" and "Secondary" Storage options configured. Also, with the above configuration - the NVMe "Pool" will have the dual role of storing the "system" Shares (appdata, domains, isos & system) on the NVMe disks and they will also be used to store new files that are saved to the "media" Share and then moved when "mover" runs. ("Mover" can also be manually run by clicking on the "Move" button at the bottom of the 'Main' GUI page.) If you do plan to wipe your F4-424 PRO, then have a go at the unRAID trial, creating various configurations (Array, Pool, Shares), copying files onto the Server then running the "Mover" feature to understand how it works.

For the current release of unRAID - 6.12.11 - one option would be: 1 x unRAID "Array" consisting of 3 x 8TB disks and 1 x 8TB Parity disk (this will provide 'protection' for the loss of up to 1 disk) 1 x unRAID "Pool" configured as a mirrored pair of the two 1TB NVMe disks The "system" Shares (appdata, domains, isos & system) should be stored on the NVMe Pool. Create a "media" Share set to Primary Storage = NVMe "Pool" and Secondary Storage = HDD "Array". Then install the appropriate Docker Applications via the "Community Applications" plug-in. Take care when configuring the correct (File) paths within Docker Applications to make sure 'downloads/data/files' are stored in the correct location (typically the NVMe "Pool" for this configuration). If you have any existing data/files/media then you will need to store these elsewhere then transfer back onto the unRAID Server once it's up and running. There is also the possibility to create a configuration using two ZFS "Pools" - 4 x 8TB in raidz1 and 2 x 1TB in a mirror - but for unRAID 6.x.y you still need to have a 'traditional' unRAID "Array" with this configuration. You can use a single, small-ish USB 'thumb drive' for this purpose (nothing will be stored on the 'thumb drive' "Array"). ZFS does typically offer much faster read and write speeds but this comes with the downside that all disks in the ZFS Pool need to be up and running to read or write a file. With the 'traditional' unRAID "Array" only the disk containing the (media) file will be spun up (if powered down) to access the file. Also, the 'traditional' unRAID "Array" supports disks of different sizes (although Parity Disk(s) need to be as large as or larger than any 'data' disk). Although, one downside of the way the 'traditional' unRAID "Array" works is that write speed is usually slower (although for a media server this is not normally a problem - write once then read many). [Don't forget to set "Settings -> Disk Settings -> Default Spin Down Delay" within unRAID to spin down disks after the configured period of inactivity.] My recommendation, with being new to unRAID, would be to go with the 'traditional' unRAID "Array" plus the mirrored "Pool". Not sure if you are aware, but you can download a trial version of unRAID, save to a USB 'thumb drive' and then test out all the features during the trial period duration. You will also need to figure out how to get into the BIOS of the F4-424 PRO and configure it to boot from "USB". 👍

One recommendation I would make (from previous personal experience) would be to just add the minimum amount of disks (plus maybe 10% extra) that provide the level of storage you currently need. It's easy enough to add additional disks to expand the unRAID Array later on whereas emptying then clearing (zeroing) disks can take some time and needs care to ensure you empty/wipe/remove the correct disk(s)!

Hi @Reynald, Thank you. I also did some investigation and saw in an old "mover_tuning" file that it was reporting that I had a Share .CFG file present that was no longer valid and that the "Mover Tuning" plugin was "skipping" this share. I deleted the Share .CFG file from the /config/shares folder and the "Mover Tuning" plugin is now working and no longer erroring (although I still have the plugin in 'test' mode at the moment). I'm not sure if this was a change in the latest version of the plugin (that I installed today) as this "missing" Share has been present for quite some time and previously did not cause the plugin to error. I have sent you a DM with the log files for the CTIME/MTIME issue I am encountering (no rush on these though). Again, thanks for your help and a great unRAID plugin!

Hi @Reynald, I wonder if it has anything to do with the fact that I have used dashes "-" in my Pool Names (e.g. my current (SSD-based) cache pool is called "os-pool" (without the double-quotes))? Is this possibly being filtered or not extracted due to this naming standard? Thanks.

Hi @Reynald, Thanks. It's a little strange as I had already set this value to 75%: And when changed to 80%, the same error persists: move: /usr/local/emhttp/plugins/ca.mover.tuning/age_mover: line 421: * 80 / 100: syntax error: operand expected (error token is "* 80 / 100") I also tried changing to 85% and 85% matching your settings but the same error occurs: move: /usr/local/emhttp/plugins/ca.mover.tuning/age_mover: line 421: * 85 / 100: syntax error: operand expected (error token is "* 85 / 100") Thanks.

Hi @Reynald, Could you advise where I set the value for the "global freeing threshold" please? Thanks.

Hello @Reynald, When attempting to run the latest version of the Mover Tuning plugin, I am receiving the following error message: /usr/local/emhttp/plugins/ca.mover.tuning/age_mover: line 421: * 75 / 100: syntax error: operand expected (error token is "* 75 / 100") unRAID then shows that "mover is running" but I have to then manually terminate by running the 'mover stop' command as it does not complete or time out. Also, I changed Mover Tuning to use CTIME (instead of MTIME) but some recently downloaded items were also being picked up to be moved when I would have expected them to be left as there were files that were downloaded many weeks earlier that I would have expected to be moved first. Diagnostics and "Mover Tuning" settings attached. Thanks. littlebox-diagnostics-20240804-1447.zip

I've used vPro on a few older generation Intel i5 processors with Intel motherboards and also HP EliteDesk 800 mini PCs. Coupled with VNC Viewer Plus (paid for edition) and once configured so that it does not require local "approval" prompts - it's a decent out-of-band solution. Although, I'm not sure about the latest version of vPro/AMT as Intel may have split it up into different feature sets. Also, my use was just for home lab and also an office based "development" PC which was occasionally switched off due to building power issues. I connected via a VPN then powered it back in using VNC Viewer Plus.

You can add multiple container paths to the Plex Docker configuration. I have one for TV Shows and one for Films/Movies. My understanding is that if you change the path for media that Plex has already indexed then it will cause it to rescan all the "new" files.

Another option might be to purchase a PCIe card that uses a PCIe switch chip (these don't require bifurcation motherboard support). I purchased a LinkReal (via Ali Express) quad NVMe card - http://www.linkreal.com.cn/en/products/LRNV9547L4IPCIExpressx16toQuad.html that can be installed in any gen3 x16 slot. All 4 NVMe/m.2 SSDs appear as separate drives within unRAID and in my case, I created a 4 drive ZFS raidz1 Pool. I also added small heatsinks to each NVMe/m.2 SSDs as they can run hot with a sustained set of file transfers.

I downsized from a Fractal Define 7 case that was just too big (and very heavy) to a Jonsbo N3 with an ASRock Z690M-ITX/ax (mITX) motherboard, i3-13100T CPU, 64GB RAM and a couple of NVMe/m.2-to-6-port-SATA cards. I also picked up a couple of dual 2.5" disk PCI Slot "adapters" and installed an additional 4 x 2.5" SATA SSDs in the two empty rear case slots (together with another couple of 2.5" SATA SSDs screwed into the existing inner case rails). Connecting everything up was a bit tricky due to limited space and I also needed to purchase some SATA-to-SATA power extension leads and some SATA-to-MOLEX adapters too. Routing the many SATA cables also requires a bit of careful organisation. Overall, I think the result is excellent, although I have used both NVMe/m.2 slots on the motherboard for SATA controllers but for my workloads, a couple of ZFS Pools (mirror & raidz1) provide the performance I need. I "could" just about utilise the x16 PCIe slot on the motherboard by moving one pair of 2.5" SATA SSDs to sit on top of the PSU (there's just enough room using double-sided sticky tape) or this space could be used to add another couple of 2.5" SATA SSDs - making the case "totally" full (8 hot-swap HDDs plus 8 SATA SSDs). I also replaced the pair of Jonsbo fans with Noctua fans and installed the "Dynamix Auto Fan Control" plugin which keeps the fans spun down unless the HDD temps start to rise. It's virtually silent with the HDDs and case fans spun down (the CPU fan does spin all the time but is very very quiet). I do agree, the N3 case only supports mITX but my understanding is that Jonsbo are bringing out the N5 case which will support 12 x hot-swap 3.5" disk bays and ITX, mITX and mATX sized motherboards too.

Ok - so nearly what I am after but not quite (I'd like to be able to add vdevs of any type and any width - I think this is technically possible with ZFS). Although, personally I would always aim to add vdevs that include redundancy (e.g. mirror, raidz(n) etc). I do understand that ZFS can be complex and that providing a simplified interface makes adoption of the technology easier but I do think a "custom" option would be useful for those that want to explicitly configure exact settings. I will give the ^ above a try though (I've an unRAID Pro license/USB stick that I use within VMware Workstation for testing out upgrades and various procedures before trying them on my "production" unRAID boxes).

Would be nice to see: Ability to take selected Pools offline to make configuration changes without needing to stop the whole "Array" & Pools (ideally keeping Docker and/or VMs running on Pools that have not been taken offline). Ability to make changes to Share settings that can then be updated without restarting the underlying SAMBA service(s). For Windows Clients, if there is an active file transfer underway when the SAMBA service(s) are restarted then it causes an error which stops the file transfer (although this can be "resumed" on Windows Clients). Not sure if it's possible to just "update" the SAMBA service(s) configuration settings rather than a complete stop and start. Ability to have an "Advanced" ZFS configuration page where "custom/explicit" ZFS Pools can be configured. At present unRAID "automagically" selects the ZFS Pool configuration based upon the number of assigned devices (e.g. 2 = mirror or stripe, 3 = mirror, stripe or raidz1, 4 = mirror, stripe, raidz1 or raidz2, etc) but when I tried to create a 9 device array with 5 x 8TB disks raidz1 and 4 x 4TB disks raidz1 this was not possible. The option to override this behaviour and explicitly configure a ZFS Pool would be good. Ability to add additional "VDev(s)" to an existing ZFS Pool. My understanding is that a ZFS Pool can be expanded by adding an additional (ideally redundant) ZFS VDev to an existing ZFS Pool. Adding this feature would also support my "Custom" configuration (Item 3.) request as I could then create a 5 disk raidz1 configuration then "add" an additional 4 disk raidz1 configuration within the same ZFS Pool. (Please note I am not talking about ZFS "RAIDZ Expansion".)

Upgraded from 6.12.10 to 6.12.11 - no issues up to now.

You could also try enabling "turbo writes" ('Settings -> Disk Settings -> Tunable (md_write_method)' to "reconstruct write"). This will slightly improve write performance to the array but will spin up all array disks as it performs write operations. Give it a try and if no difference in write performance then switch the setting back to the default. Other options that I have used within sabnzb are disabling the "direct unpack" option and enabling the "pause downloading during post processing" option to slow down sabnzb so it only process one item at a time. This can help when disk processes can't keep up. It might take longer overall but can save reaching the point where the download location is full which will stop all downloads!

Maybe try this?

The "install as App" feature on Android is pretty nice. Gives a little bit more room on screen as there's no browser bar present. 👍

For my use cases where I have both media files (Films & TV shows) and non-media files (home files/documents/photos etc) the option to have both an unRAID Array and ZFS Pools - I find to be an excellent solution. I use the Array for media files and the ZFS Pool(s) for non-media files. The Array provides the benefit of being able to use different sized disks and also easily supports disk spin down for power saving/reduced noise. The ZFS Pools for non-media files provides excellent read AND write performance and together with a ZFS special vdev (using SSDs in a 3-way mirror) also provides an experience where I can view the file & folder names of the ZFS Pool without needing to spin up the HDDs within the ZFS Pool (unless needing to read the contents of large files). I have also configured the ZFS special vdev to store "small" files which also helps with keeping the ZFS HDDs spun down unless needing to read "large" files.