I've been getting BTRFS errors on my cache for the last 2 years (ignoring them obviously doesn't work).

It first happened  two or three months apart and didn't bother me but recently it have happened as often as once a week.

 

When the errors first appeared I had two 120 GB SSDs and they are replaced by two new 250 GB SSDs. I have also changed SATA cables and tried different ports.

 

I have tried different scrub commands with various success rates, the first few times it made cache functional again but lately it says that the problem can't be fixed.

Nowadays when the errors appears I just format the drives and install dockers and run a restore from the nigthly backups.

Backups run at 3 in the morning and the server restarts every monday morning at 5.

 

My knowledge of Linux is quite basic, I know some very basic commands but that's it.

 

I hope you can help me even if the help is just a recommendation to buy a new computer/server. The current one is a bit underpowered and the kids want one or two Minecraft servers with mods  

hellplant-ur-smart-20190916-2152.zip hellplant-ur-diagnostics-20190916-1952.zip

Run a memtest and after re-format pool to start over, if there aren't RAM issues and problems persist post new diags at first sign.

Well, the test have been running for 19 minutes and is 55 % in and already found 160 errors (and counting).

 

Might be time to buy new memory.

 

Thanks for the help!

23 minutes ago, Hellplant said:

Well, the test have been running for 19 minutes and is 55 % in and already found 160 errors (and counting).

 

Might be time to buy new memory.

 

Thanks for the help!

Any value other than 0 is a problem!   Sometimes it is just the RAM needs reseating so it is worth checking that.

Edited September 17, 20196 yr by itimpi

18 minutes ago, itimpi said:

Any value other than 0 is a problem! ;(   Sometimes it is just the RAM needs reseating so it is worth checking that.

Reseated the RAM and started getting errors right att start of the test again. One minute in and 21 errors.

Try testing individual modules in different slots to see if it is a specific RAM module or a specific slot giving the errors.

On 9/17/2019 at 10:57 PM, trurl said:

Try testing individual modules in different slots to see if it is a specific RAM module or a specific slot giving the errors.

Had time to continue testing this evening and the result is four fully functional slots and one (of two) functional 4 GB RAM module.

 

Thanks to everyone for the help.

 

EDIT: I will of course monitor the cache drives and report in a week or two regardless of outcome.

Edited September 18, 20196 yr by Hellplant

As promised, here comes a update.

 

Everything works great 

 

Thanks again for the help!