plex and all my dock containers and webgui become unreachable

[Gizo]

My dockers keep failing to load on webgui and crashing and then i'll get failed to start docker services sometimes. My server was running for about 86days prior to this without much change. No dockers have even updated recently. I could only get a syslog this time before I had to reboot. I've cleared the docker.img and re added the few I truly need. But it seems to just happen all over again. When/if it does I'll try to add a diagnostics. Any help greatly appreciated, Thanks.

 

Edited: updated title and attached a couple diagnostics in latest post.

 

gizbo-syslog-20200818-1912.zip

Edited August 20, 2020 by Gizo

[Squid]

You should post the full diagnostics.  Far more info in them.

[Gizo]

2 minutes ago, Squid said:

You should post the full diagnostics.  Far more info in them.

Here is diagnostics. All I've done so far after this happen was delete the docker.img, reinstall my dockers then the same thing happened. Webgui crash's and server unreachable so I did that again. I've been up and running for about 2 hours now without any issues. Isn't it better to try and get diagnostics while the issue is happening?

 

gizbo-diagnostics-20200818-1553.zip

[trurl]

I see you did memtest several months ago. Wouldn't hurt to do it again.

[Gizo]

On 8/18/2020 at 4:44 PM, trurl said:

I see you did memtest several months ago. Wouldn't hurt to do it again.

That was on completely different hardware actually. I original setup the server using all new parts. I was getting kernel panic errors left and right and could barely use the webgui. So I ran memtest, bought all new parts one by one. Same thing happened. So I used my gaming pc and turned that into my unraid server since I knew all the parts were working in my gaming / windows machine. I was up and running great for months, only went down due to power outage. My UPS had to get sent back for repairs otherwise I'd have longer then 86 days uptime. I'm starting to think it might be network related but I'm really not too sure. Seems the server works fine for a few hours then plex and all my dock containers and webgui become unreachable. I've managed to at least get a couple diagnostics though so maybe someone will know.

gizbo-diagnostics-20200820-1251.zip gizbo-diagnostics-20200820-0619.zip

[trurl]

Unclear if you have you done memtest on this new hardware lately. The reason I suggested it is because your syslog is full of segfaults from your containers. Bad RAM is a possible cause.

 

I guess another possibility is corrupt code in docker.img. I see you have 30G for docker.img. I usually recommend 20G and often ask if there is some reason the user has set it larger, though usually only when they have set it very large. Have you had problems filling docker.img?

 

 

[Gizo]

2 hours ago, trurl said:

Unclear if you have you done memtest on this new hardware lately. The reason I suggested it is because your syslog is full of segfaults from your containers. Bad RAM is a possible cause.

 

I guess another possibility is corrupt code in docker.img. I see you have 30G for docker.img. I usually recommend 20G and often ask if there is some reason the user has set it larger, though usually only when they have set it very large. Have you had problems filling docker.img?

 

 

After the server became unresponsive the dockers stopped working. I was getting docker service failed to start. Something random I did notice when this first happen was my nvidia shield also was having trouble connecting to my network. That's why I ended up trying to view my server via webgui and got my first 500 internal server error. Restarted server and got dockers services failed to start. When I looked around on how to fix this I saw someone recommend to delete docker.img and increase the size. After doing that and reinstalling the dockers containers my server became responsive but seems to crash and becomes unreachable after a couple of hours. I can change the docker.img back to 20 at least also I'll run a memtest.