Bridged dockers become unavailable. Work again after server reboot

[winterx]

Hello.

I have a few dockers with static IPs set via br0

They work great for unspecified amount of time, 10 until the last occurence, then become unreachable.

my setup

192.168.0.1 - gateway

192.168.0.2 - server

192.168.0.10 - dock1

192.168.0.11 - dock2

 

i have console access to dock1 from unraid web interface

 

i CAN NOT ping dock1 from server . dest host unreachable

i CAN NOT ping server from dock1. dest host unreachable

i CAN ping gateway or dock2 from dock1

 

looks like by bridge gets broken but i can not find a way to prevent it or restore without reboot.

 

Any advice appreciated.

How to catch a bug?

What log can contain useful data?

How to repair bridge wo reboot?

 

 

[ljm42]

I'm guessing that on Settings -> Docker, you have "Host access to custom networks" enabled?

 

Go to Tools -> Diagnostics and download your diagnostics.

 

Then go to Settings -> Docker and disable Docker, Apply then enable Docker, Apply. Hopefully that solves it?

 

If so, grab a second set of diagnostics and upload both here. I am trying to track down what changes when disabling/enabling Docker that resolves this.

[winterx]

19 minutes ago, ljm42 said:

I'm guessing that on Settings -> Docker, you have "Host access to custom networks" enabled?

>>ofc. it works just fine for weeks

Go to Tools -> Diagnostics and download your diagnostics.

>>Done

Then go to Settings -> Docker and disable Docker, Apply then enable Docker, Apply. Hopefully that solves it?

>>That solves it, correct

If so, grab a second set of diagnostics and upload both here. I am trying to track down what changes when disabling/enabling Docker that resolves this.

>>Done

And many thanks for your effort.

 

unraid-diagnostics-20221105-0022.zip unraid-diagnostics-20221105-0026.zip

Edited November 4, 2022 by winterx

[ljm42]

Glad to hear you can recover without having to do a full reboot!

 

I'm working on other projects at the moment, would appreciate if someone can help identify differences between these two diagnostics that would explain how enabling/disabling Docker resolves this

[winterx]

5 minutes ago, ljm42 said:

Glad to hear you can recover without having to do a full reboot!

 

I'm working on other projects at the moment, would appreciate if someone can help identify differences between these two diagnostics that would explain how enabling/disabling Docker resolves this

reboot issue was on my side. i had an access through another docker so reboot was my way to restart all of them at once. wg vpn solved it.

[ich777]

On 11/4/2022 at 10:54 PM, winterx said:

reboot issue was on my side. i had an access through another docker so reboot was my way to restart all of them at once. wg vpn solved it.

Can you please give me the output from:

grep -R "" /proc/sys/net/ipv4/*

when it is working and when it's not working next time before you fix it (preferably in text files named something like "bad" & "good")?

[winterx]

On 11/7/2022 at 12:17 PM, ich777 said:

Can you please give me the output from:

grep -R "" /proc/sys/net/ipv4/*

when it is working and when it's not working next time before you fix it (preferably in text files named something like "bad" & "good")?

bad

https://pastebin.com/rdnRu0rV

good

https://pastebin.com/5v7fbiBK

 

 

i have a feeling that today it happened when i pointed a new docker to br0 stack

Edited November 13, 2022 by winterx

[ich777]

4 minutes ago, winterx said:

i have a feeling that today it happened when i pointed a new docker to br0 stack

Thank you, will look into that and report back... please give me a few days...

[winterx]

On 11/14/2022 at 12:10 AM, ich777 said:

Thank you, will look into that and report back... please give me a few days...

Hi. It looks like you never had a chance to check it out but it still happens every few days.

 

I am going to make a cron script to ping docker in question and make

/etc/rc.d/rc.docker restart

but it does not look like a permanent solution

 

update:

 

bad

https://pastebin.com/82BgJ7kS

fixed

https://pastebin.com/nxBA8zYZ

 

[ich777]

17 hours ago, winterx said:

Hi. It looks like you never had a chance to check it out but it still happens every few days.

Sorry forgot to reply, real life came in the way…

 

I see nothing obvious from the output…

 

What do you mean that it happens every few days? Do you reboot the server that often or do you loose host access randomly?

 

May I ask for what you need host access?

[winterx]

5 hours ago, ich777 said:

Sorry forgot to reply, real life came in the way…

 

I see nothing obvious from the output…

 

What do you mean that it happens every few days? Do you reboot the server that often or do you loose host access randomly?

 

May I ask for what you need host access?

NVM I totally get it. just had a hope it would be something obviuos.

 

My topic was badly formulated. Let me explain.

 

1. I have a router that routes selected IPs from lan to vpn tunnel

2. To make some of unraid containers work with this setup i gave them fixed ips with bridge network.

3. I do not monitor those dockers 24\7 but once in a week i do and sometimes they are not available

4. Rest of a system work perfectly fine.

5. Remote access to this server was configured through another container so I had to reboot because there is no "restart" option in docker settings menu and "stop" would kill my remote access as well

 

Now I have a wireguard confiured so when it happens again I can restart docker service without reboot.

My setup is not a best one ofc but it had months of uptime.

 

 

Edited December 11, 2022 by winterx