I’ve attached my Diagnosis Report. Fix Common Problems is reporting that Call Traces have been found on my server. Any assistance is appreciated. 

homeserver-diagnostics-20171009-1231.zip

 

I’ve also noticed that copying between drives in my server is super slow.  It looks like it’s copying bytes/second at the moment.  I’m able to read files off of the server no problem though. Seems to only be an issue writing to the Pool. 

Edited October 9, 20178 yr by bisk

Looks like my server was completely hung this morning. I wasn’t able to access any of my Dockers or the Web UI. I’ve attached a picture of the console before a reset it. 

This should help with the memory issues:

 

 

Thanks johnnie.black.  I read through the thread and set the cache to 50% of the defaults.

 

I’m still seeing really slow writes to my pool.  Downloading with SABnzbd to a drive outside of the pool is fast though. I ran a test with Disk Speedtest plugin and I’ve attached the results from that. All of the drives seem to be running fine, so I’m stumped on why the writing it so slow...

 

Edit - I see that that is just a test of the read speeds only, so I don’t have a view into the write speeds   Although my read speeds seem to be an issue too since I’m not able to play anything through Plex anymore either. 

Edited October 10, 20178 yr by bisk

You cache device is fully allocated, so in practice it's no being trimmed, this should help:

 

 

8 hours ago, bisk said:

I read through the thread and set the cache to 50% of the defaults.

 

This may still be too high, 1 and 2 are the recommend values.

Thanks again.  I ran 

btrfs balance start --full-balance /mnt/cache

and it said that it had to relocate 479 out of 479.  I've also changed the cache sizes to 1 and 2 as you suggested.

I had an email from the Fix Common Problems plug-in this morning stating that I had Call Traces again. I’ve attached my Diagnostics Report.

homeserver-diagnostics-20171014-1105.zip

Edited October 14, 20178 yr by bisk

Memory issues again, you've since updated to v6.4, that kernel shouldn't need the previous tweaks, try setting them back to default.

So I've since upgraded to v6.4_rc10b and I'm still receiving Out of Memory and Call Traces errors from Fix Common Problems.  My monthly Parity check started last night and was running at around 1MB/s when I checked this morning.  I've set the cache % back to 1 and 2 to help with the Out of Memory errors, but it would appear that that hasn't helped.  Any help in resolving this issue would be greatly appreciated since I'm not able to use Plex and I have some you and female family members that are constantly bugging me

 

 

homeserver-diagnostics-20171102-0720.zip

41 minutes ago, bisk said:

 I've set the cache % back to 1 and 2 to help with the Out of Memory errors, but it would appear that that hasn't helped.

 

This shouldm't be needed for v6.4, but did you reboot after applying them?

Yeah, I remember you mentioning that above, so I had set them back to the defaults, but I continued to receive the Out of Memory errors, so I put them back 1 and 2 a little while ago.  There have been many reboots since changing that setting.  I usually reboot the server after dinner if I know that we're going to watch Plex that evening since a reboot will restore the server to acceptable performance to use Plex.

Then it's probably some docker or plugin causing the issues.

OK, thanks for the help.  Short of removing all of the plugins and re-adding them one-by-one and stopping all of the dockers and starting them one-by-one each day, is there anything else I could do to help ID the troubled one?

1 hour ago, bisk said:

is there anything else I could do to help ID the troubled one?

 

Don't know other way, it's also possible it's none of them, and you need a swap file or any other memory management improvement.