itimpi

You have no shares configured to move from cache to array. For that to happen the share needs to be configured with Primary=cache, secondary=array, mover direction=cache->array. I notice the 'domains' share is configured to only have files on the 'vm' pool, but actually has some files on the 'cache' pool - is that intended? In addition the 'cache' pool has no value set for the Minimum Free Space. You want that set so that the if the 'cache' pool gets below that amount while caching a share it starts bypassing the cache and writing directly to the array. Typically you want that to be set to something like twice the size of the largest file you expect to cache.

your syslog is being spammed by messages of the form (lpcfg_do_global_parameter) Feb 21 05:50:51 Unraid smbd[2125609]: Ignoring unknown parameter "og level" Feb 21 05:51:51 Unraid smbd[2127942]: [2026/02/21 05:51:51.035651, 0] ../../lib/param/loadparm.c:750(lpcfg_map_parameter) Feb 21 05:51:51 Unraid smbd[2127942]: Unknown parameter encountered: "og level" Feb 21 05:51:51 Unraid smbd[2127942]: [2026/02/21 05:51:51.035684, 0] ../../lib/param/loadparm.c:2004 Looks like you have an invalid entry in smb-extra file.

Not normally something that you would do as the 'go' file is part of your configuration information. The default 'go' file does nothing except start up the webGUI so was probably the same on both sticks on your setup?

Not sure why you would think this clears the hardware. This is something we see frequently in hardware issues, especially if the BIOS is set to automatically start the server if power is available.

Note that there is also a Minimum Free Space setting for the 'cache' pool that is independent of the setting at the User Share level. You definitely want this set to stop the pool filling up too much.

If the emulated disk is not mountable then the rebuilt disk would not be mountable either.

You time how long it takes to stop the array tidily by pressing the Stop button in the webGUI and then provide for a margin above that. There ae separate timer settings available for closing VMs, all docker containers, and then stopping the array. so make sure they are all large enough.

have you made sure that your UPS settings are such that Unraid tries to attempt a tidy shutdown after a few minutes without power? You do not want the UPS to ever have insufficient stored power left to allow for a tidy shutdown.

It could be worth trying using a Molex->SATA splitter instead of the SATA->SATA ones to reduce the number of drives on the PSU SATA connectors.

We normally recommend against trying to split a SATA power connector more than 2 way as there can be issues getting enough current to all drives. A Molex->SATA splitter can normally split 4 way without issue.

Do you have a lost+found folder on the drive? That is where the repair process puts any files for which it could not find the directory entry to give them their original name. Not having that folder is a good sign😀

Is the SSD in the array or a pool? What format have you set it to use? Do you have Exclusive access enabled for that share? You are likely to get better informed feedback if you attach your system’s diagnostics (with everything in the one zip file) to your next post in this thread. It is always a good idea when asking questions to supply your diagnostics so we can see details of your system, how you have things configured, and the current syslog. I could probably have answered the questions I have just asked myself if the diagnostics were available.

You cannot. If your pool is redundant, then you could copy files off the pool and remove one disk. Failing that it is case of getting all files off the pool before removing any drives from the pool.

There is no standard command that I know of that gives that information about the array in an easily understood format. However if you install the parity check tuning plugin even if you do not use any of its features it does add the 'parity.check' CLI command that allows you to query the array operation status and start/stop the operation or the whole array amongst other things. I could easily enhance the information provided on status to give everything you mention although the critical stuff is already there. The current log is always at /var/log/syslog so you can query it using standard linux CLI commands. You can also view it easily via the GUI.

If you suspect a drive is faulty you should try to run the extended SMART test on the drive. Failing that will confirm that it needs replacing, whereas passing normally means the drive is OK. Earlier on it was mentioned that the drive could not even complete the Short SMART test which sounded ominous.

The /mnt/remotes location is where UD mounts shares from other machines. That command will be deleting all the files in that share on the remote machine that is mounted at that point (probably not what you want). It might upset UD in that you also delete the local mount point but surprised that had anything other than a temporary effect.

Maybe, but OP described rebuilding parity in steps 3 and 4, so I assumed steps 5 and 6 were about replacing the data drive. The steps laid out say that SATA4 will be the new parity drive and SATA1 will become a data drive. However looking more closely I do see the number of data drives is increasing from 2 to 3 so the format WILL be necessary.

There is no indication in the diagnostics that your drives have issues. Read errors are far more often due to external factors (e.g. cabling, power) than problems with the drives themselves. If a drive can complete the Extended SMART test without error then it is normally healthy.

Got Unraid version numbers slightly wrong? Not supported in current 7.2.3 Unraid release but expected to be supported in 7.3.x series?

If you want to keep the data already on the drive replaced in step 5) you must NOT format the drive in step 6 but instead let the system rebuild the drive to keep its data. If you do format in step 6 then you would have to find a way to attach the old 4TB SATA1 drive and copy its data back.

You are likely to get better informed feedback if you attach your system’s diagnostics (with everything in the one zip file) to your next post in this thread. It is always a good idea when asking questions to supply your diagnostics so we can see details of your system, how you have things configured, and the current syslog.

Just to check the obvious, have you checked the CPU fans are working properly, and also that it is not a case of dried out thermal paste on the CPUs.

Just make sure they are in their own share, and then configure that share to not use mover.

The diagnostics show that all disks are empty which suggests that somehow they were formatted or a global delete was run. You do not have your server exposed to the internet by any chance as there have been occurrences of bad actors doing this in the past. I would think your best chance of recovering any data is to see if a file recovery program like UFS Explorer on Windows can see anything. That software is not free, but it does have a free trial that would show what (if anything) it could recover.

Do you know if there are any plans to add to mover the option to Clear a disk after emptying it so that it can be removed without affecting parity? Feels like it might be a nice addition in a scenario like this one.