itimpi

Note that this allocation mode tends to be the worst from a performance perspective.

Exclusive means it is equivalent to cache-only and is optimised for performance. It is discussed in the 6.12.x release notes. There is normally no reason for the 'system' share to not also be in the same mode as that optimises performance of docker containers and that share is normally relatively small in terms of size. It makes sense for the 'syslog' share (if you have set one up) to end up on the array so I would have that set with both primary and secondary and the mover direction as pool to array.

You can replace it with a larger disk if it is no larger than the smallest parity drive. If you want to upgrade parity to a larger size and use the old parity drive as the replacement then there is the parity swap procedure for this special case.

The following should achieve what you want and give best performance: Stop the docker and VM services as they can stop files being moved Set the 'appdata' (and ideally also 'system share) so that a pool is the primary and the array is the secondary Set the mover direction to be array->pool Manually run mover to get all files moved off the array to the required pool In Settings->Global Pool Settings make sure you have the Exclusive share option set. For each share that only lives on a pool remove the secondary storage option and set the Exclusive share option. Re-enable the docker and VM services.

Have you checked that there is a licence file off the old USB stick in the config folder on the new one? That would be required for a licence transfer.

It might be worth trying iperf3 tests (single stream) between your Unraid server and your PC. Have you tried using different LAN cables and/or router ports for your Unraid server?

You can also have a problem if there is a path internal to a container that is being written to the is NOT mapped to an external location. A common example is Plex writing to a temporary folder while transcoding.

Sounds more like a cabling problem (SATA or power) causing retries to happen on the drive. You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread taken after the issue has occurred on drive 12.

Only if you told it to! we normally recommend that the Keep All Assignments option of the New Config tool is used and you then make any alterations from that position.

You can activate the syslog server to get a log that survives a reboot which might give a better idea of what exactly is initiating the shutdown. There is also the option of booting in Safe Mode which would stop the Nut plugin from loading.

Unraid does not 'install' itself in the traditional sense as it runs from RAM. As was mentioned the main Unraid system files are loaded into RAM each time you boot and the archives on the flash drive they are loaded from are automatically checked for integrity. The other thing that can load binaries is plugins, but you can select to boot in Safe Mode that suppresses installing plugins if you think that one of them could have been corrupted. Most plugins would have any files they would install provided as a .tgz file that again provided a integrity check on those files, but it is always possible a plugin installs a binary that is not corrupted, but it incompatible with the current Unraid release which was the reason that Safe Mode was introduced.

Not sure why your container should stop while a parity check is in progress but performance could be badly affected. You might want to install my Parity Check Tuning plugin so you can at least automate only running the checks in increments outside prime time. I notice that your 'system' share exists on both the cache and disk1. For best performance it should just be on the cache, and this would particularly relevant if the docker.img file is on the main array. Note that you would need to have disabled the docker (and VM) services under setting if you want mover to transfer the anything related to it to the cache; manually run mover to transfer the files; re-enable the Docker and VM services under Settings. Since you are on Unraid 6.12.3 release then for shares that exist only on the cache you can take advantage of the 'exclusive' share feature to get better performance on accessing these shares. To use this you need to: Make sure that the files for the share only exist on the cache pool and are no longer on any main array disk. Make sure the shares to use this mode have no Secondary storage set (probably currently set to the array), only the cache drive as the Primary Storage. Enable Exclusive shares under Settings->Global Share settings. Not sure if you then need to restart the array at this point. Set the Exclusive share setting for each share that should use this mode.

Not an answer to your question, but a weekly parity check seems excessive. Most users would run it monthly, or even less frequently than that.

A big question is whether you are going to want to add additional drives in the future. This is easy with the traditional Unraid array, but much harder/more restrictive with ZFS arrays.

Are you sure you have got the right cables for connecting the drive? Forward and reverse breakout cables look the same but only the correct one works. another possibility is that the drive has the 3.3v issue and needs the relevant pin taping off.

You set things up wrong If you are using a 6.12.x release then you set up the array as secondary storage and mover direction as cache->array. If using an earlier release then you set Use Cache=Yes. In both cases you stop the docker and VM services under Settings and then run mover to get files moved to the main array.

Might want to see if adding preserve case = yes to the Samba global setting to see if it speeds things up in folders with lots of files. There used to be an option for this in the GUI but it seems to have disappeared so maybe it can have undesirable side-effects.

It might help if you gave some examples of the shares and disks involved so we know what to look for. it would also be useful if you deleted from the config/shares folder on the flash drive any .cfg files that do not correspond to shares that you actually have to avoid cluttering the diagnostics with excess files.

Unraid requires an x86 based machine so the RPi would not do as it is ARM based. Do you have (or are intending to buy) something that is x86 based? Not clear from your requirements.

The easiest thing to do would be to edit the script to use the actual device id type you have and see if that then works.

You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread.

This implies that the drive serial is not being reported consistently or its size is being reported differently as this combination is what Unraid uses to identify the drive. It might be worth checking if you get the same issue if you power cycle the hardware rather than simply reboot or stop/start array.

The settings you currently have only get the server to listen for log messages. To actually get any logged you have to set one of the last two options as mentioned in the link. Normally the Mirror to flash is the easiest to use.

Have you tried the Manual install method covered here in the online documentation? Does Xen support booting a VM from the USB drive?

It is best used only when debugging mover issues.