Hi,

I am running Unraid (7.2.4) on an UGREEN NASync DXP4800 with 32 gb RAM (Crucial DDR5 RAM 32GB 4800MHz SODIMM CL40 - CT32G48C40S5).

Initially it was running fine for weeks, months. After some updates until latest it feels unstable, being stable for 1-3 weeks then crashes, now being unstable within days and having problems with docker or crashes. Now since the last reboot all docker container which where set to autostart=true are now false. Every container was stopped.

Within 1 day after reboot I received the message that docker could not start eventhough it was running before. I have xfs and did setup docker as directory nativly on an nvme.

Just did a btrfs scrub for my cache and that docker/appdata nvme. no errors found.

Since some time the logo on the right from the login screen is not being displayed. Can I check somehow the state of the flashdrive? I am really looking forward to replace it now when 7.3 is stable with the internal emmc of ugreen.

Now I downloaded the diagnostics file and found the following in the systems log (warning+error)

Apr  7 08:44:37 Tower kernel: x86/split lock detection: #AC: crashing the kernel on kernel split_locks and warning on user-space split_locks
Apr  7 08:44:37 Tower kernel: ACPI: Early table checksum verification disabled
Apr  7 08:44:37 Tower kernel: floppy0: no floppy controllers found
Apr  7 08:44:37 Tower kernel: EDAC igen6 MC0: HANDLING IBECC MEMORY ERROR
Apr  7 08:44:46 Tower mcelog: failed to prefill DIMM database from DMI data
Apr  7 08:45:26 Tower rc.local: error log : /var/log/graphql-api.log
Apr  7 08:46:23 Tower rc.samba: kill 2112
Apr  7 08:46:25 Tower rpc.statd[8808]: Failed to read /var/lib/nfs/state: Success
Apr  7 08:46:39 Tower root: error log : /var/log/graphql-api.log

From what I can see there, there is a RAM error message. Could this be the reason? I just started the live memory tester plugin but can also boot memtest86 later. Until now the live test did not show any error yet. Spec-wise the ugreen is officially able to run 16gb of ram but read through postings where a lot of people were using the crucial ram with 32 ram successfully.

Any ideas, hints what else I could check? Running the diagnosis file against ai shows only the ram problem.

Thanks for your help!

Edited April 7Apr 7 by chrneu

Enable the syslog server and post that after the next crash, together with fresh diagnostics.