Hi all

 

My setup has been running fine for the past years. Now, after installing latest stable I have been having serious issues with freezing GUI and server getting unresponsive after some hours or 1 day ish. Not even possible to SSH in.

Have observed Warnings in IO wait time in Glances when putting workload on one specific drive passed through a VM. So i have had the server running without this VM for the past days, hence the passed disk have not seen any workloads..

I have my dockers running on a separate NIC with MacVlan following Squids post on how to set this up correctly so hopefully this is not the issue.

Have been trying to google my way but I'm stuck. 

 

Currently undergoing memtest on server. Have been succesfull in finding errors in a log file from a couple of days ago, attached the errors in txt file.

 

Very pleased if someone could advise on the errors found in attached log file and if this is related to my issues. Is this a failing disk or is it the flash drive? 

The flash drive have been going strong for 4 years+ and is a SANDISK 16GB Ultra..

 

Grateful for any help

 

 

 

Edited July 15, 20232 yr by whatsmytname
Removed file

11 hours ago, whatsmytname said:

Currently undergoing memtest on server

Let that finish and and try switching to ipvlan (Settings -> Docker Settings -> Docker custom network type -> ipvlan (advanced view must be enabled, top right)).

Hi, thanks for reply.

Memtest (latest) has finished during the night, big green pass so no issues there apparantly.

I have the dockers on different VLANS and can't get them working with ipvlan..

They are on separate NIC and followed this guide:

 

Could this be a corrupted DOCKER image?

 

Just updated bios now, WS246 PRO.

Have prepared new USB flash.

 

Getting worried, have had 6 unsafe shutdowns this week..so filesystem might be taking some hits.  

C

BTRFS error (device dm-4): block=2123246174208 write time tree block corruption detected

This error usually means bad RAM or other kernel memory corruption.

Ok. Strange, since Memtest passed without issues.

 

Pulled the diagnostics, without starting array - if that could be of any help..

 

Edited July 15, 20232 yr by whatsmytname
Removed file

For now no apparent issues, the macvlan related call traces might still be causing problems.

Thank you JorgeB, much obliged.

 

So, I finally got rid of macvlan. Put my other NICS in use and set to IPVLAN.

Tried to put workload onto server by invoking mover and move some files from cache to array.

CPU io wait went up to 20% at max, throws warnings in glances - but should I be worried? System did not freeze during this transfer..

 

Was able to pull diagnostics when io wait was high if that can be used to identify anything?

 

Edited July 15, 20232 yr by whatsmytname
Removed file

Everything looks good so far, and 20% io wait is nothing to worry about, it will normally go much higher.