JorgeB

IIRC Unraid uses statfs() to get those values, please post the output of: stat -f /dev/md# Replace # with disk number.

Probably there's nothing logged but do you know at what time it happened last?

Please post the diagnostics.

It's not possible due to how parity works, it's just possible to know that it's wrong.

Most likely a controller or one of the disks, unfortunately no easy way to test unless you start swapping them.

In the syslog it's log as the disk problem, but since the SMART test passed you can rebuild on top, make sure the emulated disk is still mounting before doing it, also a good idea to swap/replace cables before to rule them out if it happens again to the same disk. There are also what look like connection issues with both cache devices, though strange that it's with both, unless they share a power splitter or something.

Shares isos is not being exported, go to Shares -> isos -> SMB Security setting to change that.

Aug 31 21:09:19 Tower emhttpd: HP_SSD_S750_1TB_HBSA42140900292_350000000000027b6 (sdb) 512 2000409264 Serial Number: HBSA42140900293 LU WWN Device Id: 5 000000 0000027b6 The problem is that both devices have the same exact wwn identifier, ending in 27b6, Unraid won't allow that, hence the "device /dev/sdb problem getting Id" errors, this would not be a problem with a HBA instead of a RAID controller, based on the model yours probably cannot be flash to IT mode, but they should work fine with an LSI HBA in IT mode or if you use the onboard SATA ports.

Other way around, regular memtest won't find errors for ECC RAM. Some strange crashing going on, since it's an Alder Lake platform update to v6.11.0-rc4, newer kernel might help if it's some compatibility issue.

Yes, but you connected disk3 to the onboard SATA, so the SATA cable must be another one, and the errors followed the disk, was the power cable the same at that time?

Go to Settings -> SMB -> SMB Extras and add: log level = 3 logging = syslog Then check the syslog, note that Samba will get very chatty so it can fill it up after a few hours.

Still pretty convinced that was from a hardware issue, but keep sing the server normally and see if there are more issues.

Eth0 should always be used for the main NIC for management and with internet access, set the NiCs for direct access to eth1, eth2, etc.

It's OK to use true HBAs, like the recommended LSI models. Pools cannot be recovered the same way, best bet for those is a file recovery util, like UFS explorer.

Use the default settings, without overclocking, should be stable at those speeds, also check that "power supply idle control" is set like it should in the link above.

You should run another one, also a good idea to post the full diags to see if there are known hardware issues.

Nothing relevant logged and the above really makes think of a hardware problem, crashing after a few minutes with the array and services stopped, don't remember ever seeing a similar case that wasn't hardware related.

Run it again without -n or nothing will be done, if it asks for -L use it.

Please post that after a crash.

See here first, if that was already done or it doesn't help enable the syslog server and post that after a crash.

With the array stopped click on disk4 and set the filesystem to xfs.

Easiest way would be to create a new "data" share then move previous movies/pictures folders (shares are just top level folders) inside the data folder disk by disk, the move will be instantaneous, you can use for example the Dynamix File Manager plugin Midnight commander (mc on the console), note that when you create a share it will only create the folder in the first available disk, so you'll need to create it (with the same case) for the other ones, not sure if the plugin has an option to create folders, if it doesn't you can do it in the console: mkdir /mnt/disk1/data mkdir /mnt/disk2/data etc The plugin moves the data inside the folders, not the folders themselves, mc will move the folders.

It really depends, I have servers with 1 or 2, others with 6 or 7.

This usually means bad RAM or other kernel corruption, start by running memtest.

I saw that, there's no driver in v6.11.0-rc4, the "are you sure?" was about the diags not being a zip file, it is for me.