itimpi

There is only 1 parity check in the steps I gave! Where do you expect the second one to fit?

The following should work I think (it assumes you have a valid parity drive which you do not mention) for all shares that have files on the cache set them to Use Cache = Yes stop docker and VM services (if you have them enabled) run the mover to move all files from cache to array. if you are not sure if you have valid parity run a parity check unassign the cache drive and the 2TB drive to be replaced start the array to commit the new assignments stop the array and assign the 3TB drive in place of the previous 2TB drive restart the array to rebuild the contents of the old 2TB drive onto the 3TB drive. stop the array Assign the old 2TB drive as a cache drive set any shares you want to end up on the cache to Use Cache = Prefer run mover to move files from array to cache restart docker and VM services (if you use them). if you do not have a parity drive then the steps from 4 onwards will need revising slightly.

I want to format some text in a Sweet Alert dialog. At the moment I have the dialog working with plain text by using a parameter of the form: text: 'some text' Looking things up on the internet it looks like it should be able to do formatting by using: html: 'some html' instead of the 'text' parameter but this does not seen to be working! I would welcome any feedback on how to achieve this or thoughts on what I might be doing wrong. Is the 'html' option not supported in the version of Sweet Alert that is included in Unraid?

I obviously need new glasses

They 6.8 RC releases have been available under the ‘next’ branch for some time now.

Leaving drives spun up 24x7 will have an impact on power, but anecdotally it reduces the chance of disk failure as the spin up/down process is one of the more demanding processes on a drive.

Have you looked at the revised text that is already present in the 6.8 RC releases? It also highlights mover behavior.

Working at the /mnt/diskX level WILL update parity appropriately as files/folders are moved around.

It is important that you continue to keep backups of important data regardless of the number of parity disks. Parity protects against disk failure, but it does not protect against other potential causes of data loss.

Parity is not good if you have made any change to the disk assignments.

Unraid does not have a ‘setup’ process in the sense most other OS have! Every time you boot the system it is unpacked from the archives on the flash drive and loaded into RAM. At the start of the boot process you are presented with a menu which allows you to select from several options (you can change which one is the default). Once you have selected an option (or time runs out and the default runs) UnRAID will start loading and various messages are displayed as the hardware is detected. At the end of that you are presented with either a console login or a web page login. your description sounds like you might just be seeing the standard boot sequence?

Yes.

What does the encrypted container look like at the Unraid file system level? If it looks like a large file then Unraid will not be aware that you are writing new files do it, but merely see it as a file that is being updated in place and the cache does not get involved (unless the container file is currently on the cache).

Do you need these modules loaded before the array is started? If so then the commands to re-install the modules need to go into the config/go file. If they are only needed after array start then a script run by the CA User Scripts plugin is the better approach (with the script scheduled to run at first array start).

It might be worth booting in non-GUI mode. Some hardware/bios/Linux kernel configurations seem to only boot in GUI mode and others only in non-GUI mode. The reason has never been identified (which is why it has not been fixed). Most users run UnRAID headlessly administering UnRAID via the network so that for such users it is not an inconvenience to just set the default boot mode to be the one that works. I also have not seen any indication that this affects whether the GPU can be passed through at the hardware level to a VM.

The current version of Unraid can be used with a Trial license which is valid for 30 days, is fully functional, and is intended to allow you to validate that UnRAID runs OK on your hardware and meets your needs. The trial licence can be extended once for a further 30 days. The version of Unraid that runs for free with a limited number of drives was v5. This is been super-ceded for many years with v6 so no new user would be expected to use it. It was limited to 32 bits and had far less functionality than current Unraid

I had wondered if even in form mode you could have a text box that included all top level xml nodes that are not part of the standard templates. A feature like that might mean one rarely needed to go into xml mode (as well as showing what custom xml nodes have been inserted).

It should be available! Does the drive being rebuilt show as being mountable? If it does not then there may be corrupt at the file system level and a repair needs running.

I think this may be harder to implement than it first seems? I do think, however that at a minimum a warning on switching from xml to form mode that changes made in cml mode may well be lost should be added?

Do not see how multiple arrays help in this use case? If you want redundancy you always going to have a minimum of 2 drives active.

You could try adding the same options to a safe mode boot to see if any of those boot options are relevant or if it a plugin causing the problem.. ideally you would make a copy of the safe mode and add the options to that so you have the standard safe mode and your amended version available.

/mnt/diskX (and /mnt/cache) refer to physical drives. /mnt/user refers to user shares, and /mnt/user/XX is a view that is an amalgamation of all the XX folders on all the physical drives. It is implemented using the Linux FUSE file system support.

My immediate suspicion would be RAM related issues. I would suggest running a memory check for something like 24 hours

Not sure why you keep referring to symbolic links for files - they are actual files and not symbolic links.

Those IOMMU groups suggest that each controller on the “USB card can individually be passed to a VM. I guess it is the a case of determining which USB port corresponds to which controller.