Hello!

I'm running an old ML310e - with 8TB Parity, 8TB/3TB/3TB Data + 4 bay 2.5" caddy, with 2* 1TB mirrored SSD Cache pool, and a 4TB SSD which I added to the Array (about 9 months ago), for faster NextCloud/Immich/Other app service.

I regularly get I/O errors when running parity check on the 4TB disk - Always 856 Read errors - yet different sectors are reported.

In general, there are no issues with the NextCloud/Immich services... but I am wondering if I'm playing with fire using the 4TB as part of the Array.

Lately, I've had all kinds of issues with my Unraid installation; with Docker and nginx being unresponsive, (Duplicate Default Server - as per known issue) and nothing working even when executed via SSH; such as 'reboot', which says it's restarting, but doesn't.

I'm not happy about the number of times I've needed to hit the power button to reboot... last night it hung trying to download the diags..

I have upgraded from 7.2.4 - 7.2.6 and it didn't help.

Today I pulled out the RAID card for the 4 bay 2.5" disks, reseated it, and the cables. Since rebooting, it seems stable again, and docker apps are fine - But yes, 856 read errors still thrown on the 4TB SSD; which will not repeat if I run the Parity check a second time.

It was running absolutely fine at my old place, but needing to move out, I migrated my server to another location - It was also running fine here; until the broadband was changed to FTTP... Then, even though the network config was the same, this is when the issues with nginx/docker started to get non-responsive, on 7.2.4.

I'm about to go away for a while and I'm just looking for stability and functional availability of Unraid - I've had more success with Tailscale, than Unraid Connect, which sometimes works and sometimes doesn't.

I appreciate this isn't particularly technical - and will upload diags if it goes all non-responsive again.. ;)

Cheers guys,

Kev

Post the diags after you get the errors.

3 hours ago, JorgeB said:

Post the diags after you get the errors.

Thanks @JorgeB

kbnas-diagnostics-20260514-1126.zip

It's not logged as a device error. Try connecting that SSD to the SATA AHCI controller instead; you can swap with one of the disks, and see if it still happens.

P.S. Note that having SSDs assigned to this array is not recommended, mostly because TRIM won't work, but it is not related to the current issues.

Update - I replaced my SAS > SATA Cabling and updated to Unraid 7.2.4 > 7.2.6 and subsequently 7.3 and the server remains stable, with a massively reduced number of I/O related read errors on the SSD during the first Parity Check .... Subsequent Parity checks have shown ZERO read errors..

I still cannot really understand how Unraid got itself all messed up and becoming non-responsive - but I reseated my RAID card, and GPU when I changed the cabling; and rebooted cleanly after the first boot. (Still using USB to boot).

I'm not sure whether using Duplicati had any impact on the Unraid server becoming non-responsive - It still seems to heavily load the CPU cores, despite me pinning two cores to that Container. (That seems to be ignored) - I've read how the CPU Scheduling could be conflicting, and it's better to leave all cores available for it. But - the error was 'Duplicate Default Server' - Which appeared to correlate with a new router being fitted - But I see there are fixes around dhcp in Unraid 7.3.1... (Although I'm not using dhcp!)

Anyway - all is fixed and my server is running and responding well, remotely, from Thailand! ;)

>>Different topic entirely but - Has CA gotten worse for failed updates/dangling images? 75% of my containers fail to update successfully, unless I manually pull an image from Terminal.

And - That new Disk Talkers plugin, is AMAZING! :)